All that Glitters is Not Gold:
Unexpected Lessons from a Slum Upgrading Program in Brazil
by
Valentina Zuin
Universita' Commerciale "Luigi Bocconi"
Laurea, International Economics and Management, 2002
Submitted to the Department of Urban Studies and Planning
In partial fulfillment of the requirements for the degree of
Master in City Planning
MASSACHUSETTS INS
OF TECHNOLOGY
at the
Massachusetts Institute of Technology
June 2005
JUN 2 8 2005
LIBRARIES
C 2005 Valentina Zuin. All rights reserved.
The author hereby grants MIT permission to reproduce
and to distribute publicly paper and electronic
copies of this thesis in whole or in part
Author
I
I
)
Department of Urban Studies and Planning
June 3, 2005
Certified by
Professor Judith Tendler
Department of Urban Studies and Planning
Thesis Supervisor
Accepted by
Dennis Frenchman
Chair, MCP Committee
Department of Urban Studies and Planning
ROTCH)
E
All that Glitters is Not Gold:
Unexpected Lessons from a Slum Upgrading Program in Brazil
by
Valentina Zuin
Submitted to the Department of Urban Studies and Planning
In partial fulfillment of the requirements for the degree of
Master in City Planning
ABSTRACT
This paper looks at the Ribeira Azul Slum Upgrading Program in Salvador de Bahia
Brazil, implemented by the development agency of the state of Bahia, CONDER, and the Italian
NGO Associazione Volontari per il Servizio Internationale (AVSI), which is recognized by the
World Bank and Cities Alliance as an exemplary project. The paper aims at understanding the
areas in which it was successful, and illustrate how - despite important shortcomings - the
project has become to be considered exemplary. The paper first shows that the project is not as
participatory as international donors believe. Moreover, it illustrates that project's sustainability
is at risk because inter-governmental political competition between the state and municipal
government and party politics prevented the inclusion of the Municipality- institution
responsible for maintenance - in project planning and implementation.
If the project did not succeed at eliciting community participation - at least in so far as
influence on project design and decision making - and is likely not to be sustained, what is this
project successful at? And how do we explain that it became well known as exemplary? I argue
that the Ribeira Azul project has been successful in the following three respects: (1) Delivering
infrastructure improvements and housing; (2) Securing financial resources; and (3) Marketing its
accomplishments.
The paper explains how, in spite of the limited community participation at least is so far
as project design and decision making - which are typically considered necessary for successful
implementation of these projects - the project built houses and infrastructure. Furthermore, the
paper shows how the close connection between AVSI and the Italian Government, and between
the project and the World Bank Task Manager, played a major role in securing sufficient
financial resources not only to complete the Ribeira Azul project, but to increase its scale to the
state-wide level. Further, the paper illustrates how AVSI's marketing strategy has been
fundamentally important to increase high level officials and workers' commitment to the project
and to make the project well known among international donors. Finally, this paper argues that
marketing as well as the performance indicators used by the World Bank and Cities Alliance
explain why this project is considered exemplary.
Thesis Supervisor: Judith Tendler
Title: Professor of Political Economy
ACKNOWLEDGMENTS
During the course of this project, many people provided support, inspiration and guidance
and made my thesis become a journey. I am deeply thankful to my advisor Professor Judith
Tendler. Her engagement in my project, insights, criticisms, and comments on many versions of
this work always challenged me and pushed me to look beyond what I was seeing. I truly
consider Judith my mentor, and meeting her has radically changed my way of thinking. She
represents for me a model of inspiration and engagement, and will always be. I am deeply
indebted to her for all the attention, encouragement, and care she gave me: they meant a lot to
me. I feel extremely lucky to have had the opportunity of working closely with her.
I am also very grateful to my readers Professor Jennifer Davis and Professor Richard
Locke. Jenna's extremely helpful suggestions, generous and attentive support and guidance made
me learn a lot throughout the process. Working with her in class and on my thesis has
represented a fundamental piece of my learning at MIT, and has opened my eyes to topics that I
have never explored before. I thank Richard for all he taught me about methodology, his
constant support, and for his help in finding a framework that could bring my findings together.
I am most appreciative to my mother and father, for accepting all my ideas, for leaving
me the freedom to pursue them over the years, and for providing me with unwavering support
and help. Thanks for the sacrifices you have made over the years to let me follow my dreams. A
special thank to my dad for being my research assistant, especially on the titles and the Italian
part of this research, and to my mum for being so giving and her constant support. Thanks also
to my sister for always making me laugh and for all her love.
MIT would have been really different without the great people that surrounded me, and
represented my family in Boston: my friends Elsie, Hiba, Erin, Ella, Carli, Roger, Claudia, and
Monica. We had a lot of fun together in these two years, and I feel very fortunate that I met you.
I truly thank Alberto for all the love and support he gave me during the beginning of my journey
at MIT, and in all the important moments. I would like to deeply thank Claudia for being always
there for me in the difficult moments. Thanks to Monica for all we learnt together. I am also
very grateful to Lynn, who helped me to go complete this journey, and to look at it in a different
and more positive way than I would have done on my own. Finally, I would like to thank
Ludovica, who is far away, but in reality always very close with her advice, care, and support.
I would also like to thank my "American Thesis Team" composed of Claudia, Anna,
Monica, Gaby, Ella, and Jigar, and my "Italian Thesis Team" composed of my dad, my mum and
Ludovica. I am particularly grateful to Ludovica, Carlo, Jigar, Roger, and Claudia for all their
help when my arm was hurting and to Jigar for his patience and for making me laugh so much.
Many people in Salvador de Bahia took me under their wing. I am extremely grateful to
Medici, Annibal, Silvio, and Rosa, not only for the time and help they gave me, but for making
me understand the "heart" of Alagados and Novos Alagados. I would also like to thank Celina
Souza. She taught me a lot about politics and gave me energy and inspiration to pursue my
research while I was in Brazil. My deepest gratefulness goes to Lucia. She received me in her
house on stilts and - always with a smile on her face - brought me around in the neighborhoods.
I would like to thank Adriana, for her support while I was in Brazil.
Finally, I want to thank Sandy, Alice, Phil, and Duncan for their precious help. This
research would have not been possible without the financial support of the Program for Human
Rights and Justice at MIT, and the Emerson Fellowship.
Table of Contents
ABSTRACT................................................................................................................................2
ACKN OW LEDGM ENTS .......................................................................................................
Table of C ontents .......................................................................................................................
3
4
GLOSSARY OF ABBREVIATION ..............................................................................................
6
Introduction .................................................................................................................................
7
M ethodology .............................................................................................................................
Possible lim itations ...................................................................................................................
15
17
Chapter 1
The Ribeira Azul Slum Upgrading Program, Salvador de Bahia, Brazil
.......... 19
1.1 Introduction.........................................................................................................................
1.2 Salvador de Bahia: an overview .....................................................................................
The political context ..........................................................................................................
The urban development of the city of Salvador: policy and projects ................................
1.3 The Ribeira Azul Slum Upgrading Program ..................................................................
Location ................................................................................................................................
History of Alagados and Novos Alagados.......................................................................
Implem enting Agencies: CONDER and AVSI.................................................................
Financial Arrangem ent......................................................................................................
Institutional Framework....................................................................................................
Governance structure of the project...................................................................................
1.4 Conclusion ..........................................................................................................................
19
19
20
21
23
27
28
30
33
35
39
41
Chapter 2
Top-down approach going hand-in-hand with construction ....................................
2.1 Introduction.........................................................................................................................
2.2 Limited community participation in project design and decision making......................
Differing Views of Participation........................................................................................
Com m unity M embers' Lack Influence on Housing Design..............................................
W hy does AV SI Play only a Lim ited Interm ediary Role ..................................................
2.3 Building Infrastructure and Housing Without Community Participation in
Housing Design....................................................................................................................
The Living Conditions of Alagados and Novos Alagados Inhabitants.............................
The role of Project Design .................................................................................................
The role of Project Implem entation .......................... .............. ......................................
43
43
44
44
46
47
50
50
51
52
2.4 Community participation in project implementation: two different successful institutional
arrangements ........................................................................................................................
54
1993-2000: "Centralized" provision of customized solutions ...........................................
55
2000 to present: "Decentralized" provision of customized solutions...............................
57
2.5 Conclusion ..........................................................................................................................
60
Chapter 3
The Politics of M aintenance..............................................................................................
3.1 Introduction.........................................................................................................................
3.2 The exclusion of the Municipality ...................................................................................
3.3 Inter-governmental political competition .......................................................................
3.4 Party politics .......................................................................................................................
3.5 Positive aspects associated with the exclusion of the M unicipality ................................
3.6 Conclusion ..........................................................................................................................
61
61
61
64
65
67
68
Chapter 4
G etting financial resources and becom ing exem plary................................................
4.1 Introduction.........................................................................................................................
4.2 Securing financial resources ............................................................................................
Inter-Italian Connections ...................................................................................................
Connections with the W orld Bank...................................................................................
4.3 M arketing ............................................................................................................................
AVSI's marketing strategy ................................................................................................
Origins of AVSI's capacity to market the project ...........................................................
Increasing high level officials and workers' commitment to the project...........................
4.4 Becoming an "exemplary" project...................................................................................
The role of AVSI's marketing strategy ............................................................................
The role of the set of indicators of success used..............................................................
4.5 Conclusion ..........................................................................................................................
69
69
69
70
72
74
75
78
80
82
82
83
84
Conclusions ...............................................................................................................................
86
APPENDIX I: Neighborhoods of Alagados and Novos Alagados............................................
91
APPENDIX II: Comparison between Novos Alagados before and after the intervention..... 92
APPENDIX III: Houses on stilts ..............................................................................................
93
APPENDIX IV: Houses before and after the intervention .......................................................
95
Bibliography .................................................................................................................................
96
GLOSSARY OF ABBREVIATION
ACM - Antonio Carlos Magalhies
AMESA - Alagados Melhoramentos (1974-1983)
ARENA - National Renovating Alliance
AVSI - Associazione Volontari per il Servizio Internazionale
BNH - Banco Nacional de Habitacao
CA - Cities Alliance, Washington DC
CDO - Compagnia delle Opere
CEF - Caixa Economica Federal, substituted BNH in 1986
CL - Comunione e Liberazione
CONDER - Metropolitan Agency of the State of Bahia (1974-1998), Urban Development
Agency of the State of Bahia
HAMESA - Habitacao Melhoramentos, Housing State Agency
IADB - Inter-American Development Bank, Washington DC
MAE - Italian Minister for Foreign Affairs
PATS - Proiecto de Apoyo Tecnico e Social, Technical and Social Assistance Project
PFL - Party of the Liberal Front
PMU - Program Management Unit of PATS
PSDB - Party of Brazilian Social Democracy
SEDUR - Secretariat for Urban Development
URBIS - Habitacao e Urbanizacao do Estado, Housing and Urbanization Agency
WB - The World Bank, Washington DC
Introduction
In 2001, about 32% of the global population (924 million people), lived in slums
worldwide, without access to adequate housing, water, or sanitation facilities. During the next
30 years, urban growth will bring a further 2 billion people into the cities of the developing
world, doubling their size (UN-Habitat 2004). Providing shelter and basic infrastructure for all
city-dwellers will become an increasingly serious challenge for policy makers (UN-Habitat
2004). In the last three decades, many approaches have been attempted to improve the
conditions of slums dwellers in developing country cities, ranging from eviction to site and
services, urban management, and, more recently, slum upgrading. Yet not many projects have
succeeded at improving the living conditions of slum dwellers.
This paper looks at the Ribeira Azul program, recognized by the World Bank, Cities
Alliance, and the Italian Government as an exemplary slum upgrading project. The paper aims
to understand why this project is successful, given that many other similar projects often fail.
Started in 1993, this $60 million project - financed among others, by the World Bank, the InterAmerican Development Bank, the Caixa Economica Federal, and the Italian Government - aims
at upgrading a squatter settlement of 150,000 people living in houses on stilts in Salvador de
Bahia, Brazil. Implemented by the urban development agency (CONDER) of the state of Bahia,
and the Italian NGO Associazione Volontari per il Servizio Internazionale (AVSI), the World
Bank and Cities Alliance consider the Ribeira Azul project exemplary in five respects (Cities
Alliance 2004, personal interviews). First, unlike many slum upgrading projects that fail to
benefit a large number of slum dwellers, the Ribeira Azul project was successfully scaled up
from an initial pilot project that targeted 15,000 people to a project that today targets
approximately 150,000 people.
Second, the project provided new housing and infrastructure to a large number of people,
despite the major challenges posed by the tidal terrain of Ribeira Azul where approximately
3,500 families live in stilt housing structures. The project undertakes the difficult job of
"upgrading" the existing stilt housing structures by "filling-in" the area between the house floors
and underwater earth. Such land-fill is expensive and complex to design from the engineering
standpoint. Once the stilt houses are finally resting on dry ground, the old structures are replaced
with new houses. The project has already replaced 1,700 stilt houses and is currently replacing
an additional 600 houses. Furthermore, the project has built new housing for more than 200
families who have been displaced by infrastructure improvements. New construction of
infrastructure, such as water, drainage and roads, has been achieved even in areas where
substantial land-fill was required.
Today, approximately 7,700 families have access to new
water, sanitation and drainage systems. Between 1999 and 2005, legal water and sanitation
connections in the target areas increased almost three fold, and electricity coverage reached 90%
(see table 1).1
Table 1: Water, sanitation and electricity coverage 1999 and 2005
1999
2005
Water
30%
79%
Sanitation
29%
75%
Electricity
65%
90%
Source: Relatorio de Atividades do PATS, 2005
This figure corresponds to the areas where the project has been already concluded.
Third, the project is exemplary because of its integrated approach, which consists of the
provision of social activities beyond the infrastructure and housing improvements. In fact, the
project invested an unusually significant amount of project funding - 10% ($6 million) of the
total funding - into social activities such as training courses (for community organizations,
teachers, and illiterate women), cooperatives, and environmental and education talks (on
domestic violence, nutrition, and safety). Additionally, the project financed small physical
improvements to facilities in the areas, such as schools, health centers, sport centers, and
community organizations.
Fourth, the project succeeded at eliciting community participation, and establishing a
close dialogue and collaboration with community members and community organizations. The
participatory approach of the project is also reflected by the participation of five community
leaders of the Ribeira Azul area in the Consultative Committee. This committee is responsible
for providing guidance to the Program Management Unit for the implementation of the Ribeira
Azul Program.
Finally, the Ribeira Azul is considered exemplary because, unlike many projects that do
not succeed at creating synergies among different stakeholders, the Ribeira Azul project
succeeded in fostering collaboration among a large number of local and international actors, such
as the State Government of Bahia, the Municipality of Salvador, Community organizations, the
Italian NGO AVSI, the Italian Government, and the World Bank. All these stakeholders through their involvement in the Consultative Committee -- monitor the progress of the
implementation of the Ribeira Azul project, and provide suggestions for improvement to the
Program Management Unit.
Despite the World Bank and Cities Alliance's strong support for the project and positive
assessment of the five aforementioned areas, when I began to examine the project more closely, I
discovered that the project was deficient in two areas important areas. Contrary to what
international donors and project stakeholders believe, community participation in the project is
limited, particularly in project design and decision making -- two activities in which community
participation is considered to be essential for the success of the project. For example,
community members complain that their suggestions on how to organize the project are seldom
taken into account, and that meetings between the project staff and the community are only
informative, as decisions have already been made by CONDER and AVSI. The fact that this
NGO-led project has been "successful" despite limited participation contradicts two current
widespread beliefs in the field of development. First, slum upgrading projects are said to be
successful at delivering infrastructure and housing improvements only when beneficiaries
participate in project design and decision making. In addition, NGOs are considered the most
suited actor to ensure community participation because they understand and respond to local
needs. There is thus a need to review the project in more detail in order to determine, first, how
the project was able to succeed despite the limited participation, and, second, why there was such
little participation despite the fact that an NGO was involved in the project.
The other area of deficiency for the project was in the area of maintenance and
sustainability. As with many other infrastructure projects, my study found identified that
sustainability represents a challenge also in the Ribeira Azul project because the Municipality,
which is responsible for maintenance, has been defacto excluded from project implementation.
Such exclusion has to be explained on political grounds: inter-governmental political
competition between the State Government and the Municipality resulted in the exclusion of the
Municipality when the project started in 1993. Surprisingly, even since the state government and
the Municipality have been controlled by the same party - the PFL - the Municipality has not
been included in the project. This is the result of a decision of PFL high level politicians not to
concentrate the financial resources available to the Municipality in an area of the city already
under the political influence of the PFL - like the Ribeira Azul area - and, instead, extend the
PFL political influence over diffident areas of the city of Salvador. Moreover, the exclusion of
the Municipality is the result of personal rivalry between the leader of the PFL party - who
supported the Ribeira Azul project since its inception - and the Mayor of Salvador.
If the project did not succeed at fostering community participation - at least in so far as
influence on project design and decision making - and is likely not to be sustained, what is this
project successful in? And how do we explain that it became known worldwide as an exemplary
project?
This paper argues that the Ribeira Azul project has been successful in the following three
respects: (1) obtaining community buy-in for the infrastructure improvements and housing; (2)
Securing financial resources to scale up; and (3) Marketing-locally and internationally-its
methodology and accomplishments. I now discuss each of them in turn.
Obtaining community buy-in for infrastructureimprovement and housing
This paper explains how - despite the limited community participation in project design the project succeeded at delivering infrastructure and housing improvements. I argue that the
inhabitants agreed to resettle or to have their stilt house to be substituted for four reasons: (1)
their very poor living conditions; (2) the increasing costs of maintaining the houses on stilts; (3)
the state agency CONDER's decision not to collect payments for the houses; and (4) engineers'
willingness to slightly modify project design and provide individualized solutions to satisfy
community members' concerns.
In addition, this paper challenges the beliefs that infrastructure agencies are unable to
customize solutions to community needs and that NGOs are necessary to intermediate between
engineers and community members for successful implementation of housing and infrastructure
improvements (Nance 2004). Indeed, this paper suggests that a state agency, such as CONDER,
can also succeed in providing individualize solutions that respond to local needs. Indeed,
socially-minded engineers and architects working in a field office in the Ribeira Azul area
responded flexibly to community members' requests for small changes to project design.
Furthermore, this paper attempts to understand under what conditions engineers and
social workers successfully collaborate. While social intermediation specialists are often
requested to play an intermediary role between the community and engineers responsible for
construction, engineers often consider them low status and do not take into account their
technical input. However, the paper shows that, like in other successful projects (Davis et al.
200 1), engineers worked in close contact with social workers because social intermediation
specialists made their work easier. Additionally, informal conversation over lunch breaks
stimulated dialogue between social intermediation specialists and engineers, thus increasing
appreciation for each other's role and ideas, and collaboration among different actors.
Securingfinancialresources to scale up
International donors and scholars often cite lack of financial resources as a reason for the
lack of sustainability among slum upgrading initiatives, many of which are delayed or benefit
only a limited number of slum dwellers (Davis 2002). However, the Ribeira Azul program
succeeded in securing sufficient financial resources not only to complete the project, but to
increase its scale to the state-wide level. This paper shows how the project succeeded at
receiving a $5 million grant of the Italian Government because of the close personal connection
between the Italian Government and AVSI. In addition, this paper explains how the close
involvement of the current World Bank task manager of the Ribeira Azul program in the
preparatory phase of the project financed by the Italian Government created a solid link between
the World Bank and CONDER and AVSI. Such link resulted in the World Bank's approval of a
new $96 million loan to finance a $160 million slum upgrading initiative that will include 16
Municipalities of the State of Bahia, using the same methodology of the Ribeira Azul project,
that will start in 2006.
Marketing-locallyand internationally- the project's methodology and
accomplishments.
Like other projects that international donors and scholars have considered "Best
Practices" - such as the Indore Networking projects (Verma 2000) - the Ribeira Azul project has
been very successful in marketing itself. This paper shows how AVSI's marketing strategy
targeting international donors has been fundamentally important to the project in two respects.
First, it increased high level officials and workers' commitment to the project. Indeed, while
governments often do not find upgrading attractive for political display (Otiso 2003), AVSI's
marketing strategy succeeded at strengthening the interest of high level officials' in the project.
Indeed, high level officials started seeing the proj ect as an opportunity to enter an international
network when AVSI started presenting the projects to different international conferences,
inviting politicians and international donors to visit the project in Salvador, and inviting Bahian
politicians to present the project in Italy. Second, marketing made international donors aware of
the project, thus contributing to its fame and to make it exemplary.
Finally, this paper explores why the World Bank and Cities Alliance consider this project
as exemplary, despite its shortcomings. As we have seen, marketing has contributed to make the
project well known and advertise its accomplishments. In addition, this paper argues that the
World Bank and Cities Alliance consider this project exemplary because it conforms to their set
of indicators of success - which includes indicators such as the "ability to scale up" and "ability
to build infrastructure."
In conclusion, this paper analyzes an exemplary slum upgrading project to understand the
areas in which it was successful, and illustrate how - despite important shortcomings - the
project has become to be considered exemplary. First, the paper describes how - despite limited
involvement of the community at least in so far as project design and decision making - the
project succeeds at improving housing and infrastructure. Specifically, it answers the following
questions: How did construction happen with limited community participation in project design
and decision making? Is the presence of an NGO necessary for making infrastructure and
housing improvements in slums possible? Under what conditions social workers and engineers
collaborate to provide the community with customized solutions?
Furthermore, this paper describes how politics - namely inter-governmental political
competition between the state and the Municipality, and party politics - influence the possibility
of the project to be maintained, and therefore sustainable. More specifically, after looking at
political competition between the state and municipal levels of government, the paper addresses
the following question: why was the Municipality still excluded from project implementation,
after the state and municipal government came to be controlled by the same party?
Moreover, the paper presents other two activities in which the project is successful,
namely securing financial resources and marketing. Specifically, it addresses the following
questions: How was the project able to secure financial resources? Why and in what ways was
marketing important to the success of the project? Finally, the paper questions why - despite its
shortcomings - the Ribeira Azul project has become to be considered an exemplary slum
upgrading project.
Chapter 1 presents the political and economic context in which the Ribeira Azul program
developed, and provides an overview of the history, goals, outcomes, financial and institutional
arrangement of the Ribeira Azul Program. Chapter 2 delves into the issue of community
participation, and shows how despite the lack of community members' involvement in project
design, the project succeeds at building housing and infrastructure. Chapter 3 describes the
influence of politics on the maintenance of infrastructure. Chapter 4 illustrates what the project
is successful at, and explain how it became known as an exemplary slum upgrading program.
Methodology
This study is based on three months of field work in Salvador de Bahia, Brazil between
June and August of 2004, and a month field work in Milan and Rome, Italy, Washington, DC
and New York City in January 2005. My decision to study this project was inspired by a oneday visit to the project in the month of January of 2004, when I undertook a tour by boat of the
project area.
I spent the first month of my research in Salvador doing participant observation of the
Ribeira Azul project staff, interacting with them and with the community in the headquarters, in
the three field offices in Alagados and Novos Alagados, and during transect walks through the
communities. In the remaining two months, I interviewed 83 people, including the present and
past project staff of both CONDER (20) and AVSI (21), ten community leaders, 17 community
members (some who benefited and some who did not benefited by the project), three local
academics, and ten officials of other government agencies. Finally, I interviewed the two World
Bank officials involved in the project.
I selected the interviewees according to the position they hold or used to hold within
CONDER and AVSI, or in the project. I relied on key informants to identify officials in AVSI,
CONDER, or other institutions that, despite not being directly involved in the project, could have
important information to share about the project, or the political and economic context.
Interviews took place in the offices, while walking in the settlements, or in more informal
settings like bars, restaurants, or in the car. Some interviewees were interviewed more than once
in different contexts, in some cases also in the presence of another colleague of theirs, with the
purpose of generating discussion among them.
During my stay in Salvador, I also participated in 3 seminars, two community meetings,
two community leaders' meetings, and lunches where the project was discussed. In addition, I
attended events like an internal meeting of CONDER's staff, or political meetings where key
actors involved in the local political arena, or in the project were present. Moreover, I spent
weekends with some inhabitants of the community of Alagados and Novos Alagados to
participate in some activities they organized, or in political meetings. I visited other CONDER
and AVSI's projects, and I participated in events organized in occasion of the official visits of
the Italian Ambassador, the World Bank Task Manager, the president of AVSI, and a visit of the
Italian cooperation.
In Italy, I interviewed AVSI's management in Milan, and an official of the Italian
Government in Rome. In Washington DC, I met a total of 10 high level officials working for the
Inter-American Development Bank and of the World Bank that had been involved in the Ribeira
Azul Project or that are responsible for other urban upgrading initiatives implemented in Brazil
and other countries. Finally, I visited Professor Micheal Cohen, an urban development expert
that used to be head of the urban development unit at the World Bank, in New York City.
My interviews were conducted in Portuguese, English, or Italian, depending on the
nationality of the interviewees. I took notes whenever appropriate. In some cases I only listened
and wrote down my impressions after the interview was over, or at the end of the day.
While I was in Salvador, I collected literature on the economy, politics, and housing
policies of the State of Bahia and the Municipality of Salvador. While I was in Italy, I collected
information about the religious movement and the entrepreneur's association AVSI belongs to.
In Washington DC, I collected project documents, evaluations and discussion papers on urban
upgrading projects.
Possible limitations
The methodology for this study has a number of possible limitations. First, the fact that I
am Italian might have influenced people's opinion about the Italian NGO AVSI and in general
about the project, as they might have felt that I was associated with AVSI. I think this is a risk
of both the interviewees I was able to contact through AVSI, and those that I found through other
key informants. In some occasions I assessed that people, particularly high level CONDER's
officials and lower level AVSI's workers, were afraid that their answers or critics towards AVSI
could jeopardize their position or put their job at risk, although I guaranteed them anonymity.
Despite the fact that nobody refused to be interviewed, I had to dedicate a lot of time to build
trust with the interviewees, and I was not always able to do so. Consequently, I collected more
information through those interviewees with whom I was able to build greater trust and who
sympathized with my efforts to understand how and why the project worked well, and did not
work well.
Second, as it was dangerous to walk around on my own in the settlements of Alagados
and Novos Alagados, I was able to interview only a limited number of community members.
Only after I established personal contacts with community members by participating in activities
that the project organized for them, was I able to organize personal interviews on my own in the
community.
Another significant challenge I faced was the difficulty of obtaining project documents
and quantitative information about the amount invested in the different areas of the project, the
activities financed, and the donors financing each activity in each area. The documents I have
contain data that are not complete or do not match up. In addition, the information available on
project outcomes do not allow a deep understanding of the physical improvements made by the
project.
Chapter 1
The Ribeira Azul Slum Upgrading Program, Salvador de
Bahia, Brazil
1.1 Introduction
This chapter provides an overview of the political context in which the Ribeira Azul Program
developed and describes the Ribeira Azul Program. Section 1.2 illustrates the political and urban
development of the City of Salvador. Section 1.3 presents the Ribeira Azul Program: its location,
the history of the Ribeira Azul area, its implementing agencies, its financial arrangement, its
outcome, its institutional arrangement, and its governance structure.
1.2 Salvador de Bahia: an overview
Salvador is the capital of the State of Bahia, in the Northeast of Brazil (figure 1).
Although it has the sixth largest state economy of the country, the State of Bahia is one of the
poorest of Brazil: the annual GDP per capita is $1,562, 59% of the average per capita GDP in
Brazil, functional illiteracy rate is 41%, 58% of the population lives on less than one minimum
wage, and 25% of the total income is concentrated in 2% of the population (SEI 2003).
The political context
It is worth taking a look at the political context of Bahia and Salvador because it deeply
influenced the institutional arrangement and the outcome of the Ribeira Azul project. The
political scene of the State of Bahia and of the City of Salvador have been dominated by
Antonios Carlos Magalhaes, known as ACM (Souza 1997). ACM was Mayor of Salvador, three
times Governor of the State of Bahia, Minister of Telecommunications, and twice Senator since
1967. The modernization and economic development of the State of Bahia and Salvador is
attributed to its political leadership and even when he was not in power, ACM controlled the
Bahian economic scene and exercised political influence over private and public institutions
(Souza 1997). For example, he occupied in person or through trusted collaborators the
presidency of major companies such as Petrobas 2 , and he influenced the judiciary and
telecommunications (Souza 1997).
ACM modernized the old city of Salvador when he was mayor (1967-1970), opening
avenues in the valleys, and demolishing traditional but decayed sites. During his first term as
State Governor (1971-1975), ACM built infrastructure to support the development of the
petrochemical industry at the outskirts of the city through the Metropolitan Development Agency
CONDER. 3 This agency, which in 1998 became a state development agency responsible for
urban development and housing, played a major role in the development of infrastructure and
housing in the city of Salvador in the coming years, including in the development of the project
analyzed in this paper. Furthermore, ACM started modernizing the Bahian bureaucracy by
giving young professionals major positions in the administration, thus "creating a group of
technocrats-turned-politicians (Souza 1997: 125)." In his second governor term (1979-1982),
ACM provided electricity to the west of Bahia, improved roads and highways as well as other
infrastructure in Salvador. In his third term (1991-1994), ACM promoted Salvador's tourism, by
investing in historical sites. Summing up, the economic and urban development of the State of
Bahia and Salvador cannot be disentangled from the person of ACM.
The urban development of the city of Salvador: policy and projects
The rapid industrialization of the city, with the establishment in the late 1960s of the first
industrial parks and the petrochemical industry, resulted in the population increase and
expansion of the city in a fast and disordered manner. Many informal settlements, formed
through a sequence of illegal occupations called "invasions," grew around the city." Just
The company has become the country's leader in the distribution of oil products and is internationally
acknowledged as one of the fifteenth largest oil companies in the world today.
3 CONDER stands for Compahia de Desenvolvimento da Regiao Metropolitana de Salvador and was established by
ACM in 1974.
2
between 1980 and 1989, as many as 240 invasions took place (Teixera 2002). In response to this
demographic pressure, the state government - through its housing agency Habitagio
Melhoramentos (HAMESA), incorporated in 1987 by the new housing agency Habitagao e
Urbanizagio do Estado (URBIS) - built many housing complexes in the outskirts of the city or
near the industrial parks, with financing from the federal government. 4 However, the state
government did not succeed at providing housing for the poor: by 1986, URBIS produced 43,850
houses, of which only 2,600 (6%) were affordable for families living with 1 to 3 minimum
salaries (Gordilho 2000: 135).
In 1985, the new Brazilian Constitution decentralized the implementation of housing
policies from the State Level to the Municipal level. Still, the state agency URBIS remained
responsible for the implementation of housing policies in the following decade, but, both at the
Federal and State level, limited policy making occurred in the housing sector. This is because
the Banco Nacional de Habita9ao (BNH) 5 was dismantled in 1986 and substituted with the Caixa
Economica Federal,6 agency with limited experience and funding to promote housing policies.
Only at the beginning of the 1990s a number of different programs implemented at the
Metropolitan, the State and Municipal level started addressing the problem of providing housing
to the poor in the city of Salvador. At the beginning of the 1990s the Metropolitan Development
Agency CONDER promoted the Metropolitan Project, a $77 million program financed by the
World Bank. This project - known as one of the most important programs ever implemented in
The Federal government provided funding for the development of housing through the banks Sistema Financiero
de Habitacao (SFH) and the Banco Nacional de Habitagao (BNH).
' The Banco Nacional de Habitagao was a federal bank established with the purpose of financing housing
development especially for middle and low income population.
6 Caixa Economica Federal is the largest public bank of Latin America, and is the most important actor of the social
policy of the Federal Government. With the goal of improving the quality of life of the lower income people, the
portfolio of investments include mainly investments in the housing, sanitation, infrastructure and service sectors.
4
Bahia - brought about major improvements to infrastructure and municipal services in different
areas of the city of Salvador, including the poor area of Novos Alagados in the outskirt of the
city. The Environmental and Social Upgrading project targeting the poor area of Novos
Alagados represented the first of a number of projects that CONDER implemented in the
subsequent years to improve the living conditions of the poor. In fact, in the 1990s, CONDER
has been involved in many of such programs programs8 , such as Habitar Brazil9 , Viver Melhor,
Pro-Saneamento 0 , and Pro-Moradia.
Started in 1993, the Environmental and Social Upgrading project in Novos Alagados
represented also the first of the projects that from 1998 became part of the Ribeira Azul Program
implemented by CONDER in the area of Novos Alagados and Alagados. The Ribeira Azul slum
upgrading program is the focus of this paper.
1.3 The Ribeira Azul Slum Upgrading Program
The Ribeira Azul slum upgrading program is a $60 million project and aims at upgrading
a squatter settlement of 150,000 people (5% of the city population), where 3,500 families live in
7 For example, the Metropolitan project supported infrastructure projects such as waste water management projects,
solid waste collection, and disposal projects.
8 These programs are all financed by the Caixa Economica Federal.
9 Habitar Brazil is a US$ 417 program of the Federal Goverment, partially financed by the Inter-American
Development Bank, that is implemented through the Caixa Economica Federal.
The objective of the program is to support the federal government's effort to: i) raise living standards for poor
families living in substandard neighborhoods in metropolitan areas and urban and capital agglomerations;
ii) encourage municipal governments to allocate greater resources to alleviating both the effects and the principal
causes of problems in these neighborhoods; and iii) increase knowledge of the sectoral housing and urban
infrastructure problems in the country.
10Pr6-Saneamento is a program of the Federal Government that aims at improving the health condition and quality
of life of the poor providing greater access to sanitation.
" Pro-Moradia channels housing finance from the Caixa Economica Federal to states and municipalities (at
subsidized rates) to develop specific projects, including sites and services, construction materials packages,
construction projects, infrastructure provision in existing housing projects; and urbanization offavelas and informal
settlements. In Bahia the program is managed by the State Government of Bahia. Created in 1995, the program
targets households earning below 3 minimum salaries. Counterpart funds are a minimum of 10 percent of the total
project costs. However ,most municipalities especially the smaller ones do not have sufficient debt capacity, hence
they are not eligible for these funds.
houses stilts over a tidal zone, in the northwestern outskirts of Salvador. Implemented by the
urban development agency (CONDER) of the state of Bahia, and the Italian NGO Associazione
Volontari per il Servizio Internazionale (AVSI), the World Bank and Cities Alliance consider the
Ribeira Azul project exemplary in five respects (Cities Alliance 2004, personal interviews).
First, unlike many slum upgrading projects that fail to benefit a large number of slum dwellers,
the Ribeira Azul project was successfully scaled up from an initial pilot project that targeted
15,000 people to a project that today targets approximately 150,000 people.
Second, the project provided new housing and infrastructure to a large number of people,
despite the major challenges posed by the tidal terrain of Ribeira Azul where approximately
3,500 families live in stilt housing structures. The project undertakes the difficult job of
"upgrading" the existing stilt housing structures by "filling-in" the area between the house floors
and underwater earth. Such land-fill is expensive and complex to design from the engineering
standpoint. Once the stilt houses are finally resting on dry ground, the old structures are replaced
with new houses. The project has already replaced 1,700 stilt houses and is currently replacing
an additional 600 houses. Furthermore, the project has built new housing for more than 200
families who have been displaced by infrastructure improvements. New construction of
infrastructure, such as water, drainage and roads, has been achieved even in areas where
substantial land-fill was required.
Table 1 provides an overview of the project outcome for
what regards infrastructure and housing.
Table 1: Infrastructure and Housing built by the project (2001 and 2004)
Infrastructure
Housing
8,450,057* m3 of
1,700** (48.5%) stilt
houses removed and
substituted with new
land-fill
houses - other 800 are
currently being
removed
23,432* m of drainage
34,679* m of water
pipes
51,888* m of sanitation
pipes
44,844* m of street
246** families
benefited from housing
improvements (mainly
bathrooms)
paved
7,651 * families
136* families were
benefited from improved resettled
infrastructure
I
_I
Source: PATS (2001), Relatorio de Atividades do PATS, 2005
* 2001 data
** 2004 data
Today, approximately 7,700 families have access to new water, sanitation and drainage
systems. Between 1999 and 2005, legal water and sanitation connections in the target areas
increased almost three fold, and electricity coverage reached 90% (see table 1).12
Table 1: Water, sanitation and electricity coverage 1999 and 2005
1999
2005
Water
30%
79%
Sanitation
29%
75%
Electricity
65%
90%
Source: Relatorio de Atividades do PATS, 2005
12
This figure corresponds to the areas where the project has been already concluded.
Third, the project is exemplary because of its integrated approach, which consists of the
provision of social activities beyond the infrastructure and housing improvements. In fact, the
project invested an unusually significant amount of project funding - 10% ($6 million) of the
total funding - into social activities such as training courses (for community organizations,
teachers, and illiterate women), cooperatives, and environmental and education talks (on
domestic violence, nutrition, and safety). Additionally, the project financed small physical
improvements to facilities in the areas, such as schools, health centers, sport centers, and
community organizations. Table 3 presents summary of the social activity implemented in the
area.
Table 3: Social Activities (2004)
Social Component
42 training courses
11 social equipments (schools, community centers) built or restored
46 educational activities developed (community theater activities,
ecological walks, talks about violence, malnutrition, and hygiene
practices)
4 cooperatives and 3 commercial activities started
102 leaders of community organizations trained
10 projects directly implemented by community organizations
other activities with community organizations (including courses to
improve budgeting and management)
Source: Relatorio de Atividades do PATS, 2005
Fourth, the project succeeded at eliciting community participation, and at establishing a
close dialogue and collaboration with community members and community organizations. The
participatory approach of the project is also reflected by the participation of five community
leaders of the Ribeira Azul area in the Consultative Committee. This committee is responsible
for providing guidance to the Program Management Unit for the implementation of the Ribeira
Azul Program.
Finally, the Ribeira Azul is considered exemplary because, unlike many projects that do
not succeed at creating synergies among different stakeholders, the Ribeira Azul project
succeeded in fostering collaboration among a large number of local and international actors, such
as the State Government of Bahia, the Municipality of Salvador, Community organizations, the
Italian NGO AVSI, the Italian Government, and the World Bank. All these stakeholders through their involvement in the Consultative Committee -- monitor the progress of the
implementation of the Ribeira Azul project, and provide suggestions for improvement to the
Program Management Unit.
Location
The two areas that benefit from the program are known as Novos Alagados and
Alagados (see figure 2 and Appendix I).
Figure 2: Salvador and Novos Alagados and Alagados
Source: http://www.bahia-online.net/cityandbay.gif
Alagados and Novos Alagados are among the poorest areas of Salvador: 3500 families
lived on stilts at the beginning of the project, 45% of the economically active population does not
receive any income, 59% of the families in the area live with half minimum salary per month
($45), 13% of the houses on dry ground are not provided with sanitation facilities, and 64% of
the sanitation facilities installed are not properly functioning (Teixera 2002). Novos Alagados
and Alagados consist of a collection of bayside neighborhoods with an area of about 4km2 .
These neighborhoods, shown in Appendix I, are named as follows: Novos Alagados I Etapa,
Novos Alagados II Etapa, Novos Alagados III Etapa, and Alagados I, II, III, IV, V, VI,
Mangueira, Joanes Azul, and Joanes Centro-Oeste (see Appendix I).
In August 2004, the
upgrading had been completed in the following areas:' 3 Arajas I and II, Mudanca, Magueira II,
Novos Alagados I Etapa, Alagados I, Alagados II, and Baixa do Caranguejo. The project is still
on-going in the following areas: Novos Alagados II etapa, Alagados III, Alagados IV and V,
Joanes Azul, and Mirante Joanes Centro-Oeste. In August 2004, the State government was
looking for $10 million to extend the project to Alagados VI and Mangueira, finish the
infrastructure in Novos Alagados II Etapa and build the remaining parts to connect the coastal
road among the different areas of the project.
History of Alagados and Novos Alagados
Alagados and Novos Alagados formed as a result of a number of invasions that occurred
in the adjacent peninsula Itapagipe starting in the 1940s, when the peninsula became the
industrial center of Bahia with the development of the petrochemical industry. Alagados and
Novos Alagados are situated along the cost, in the Bay of Cabrito and in the Bay of Tainheiro.
13
Unfortunately, I did not have access to data about the percentage of upgrading completed over th total
The proximity to employment opportunities, and the possibility to avoid eviction by building stilt
houses in the tidal zones, made Alagados attractive to many squatters, although the area was
used by the Municipality as city waste disposal.' 4 Most of the squatters progressively occupied
the tidal zones and started building their stilts using waste as embankment. In some cases, after
years, the inhabitants were able to land-fill their stilts with waste, so that it could lie on dry
ground. Alagados became the biggest slum on stilts in Latin America. Many poets and artists
have celebrated the uniqueness and charm of the place. At the same time, because streets and
houses were built over the trash, and inhabitants did not have access to water and sanitation
facilities, Alagados represented a shame for Bahian politicians since the 1980s.'1
Figure 3: Life in Alagados
The uniqueness and very poor living conditions of the inhabitants, visible to anyone
visiting the place, made Alagados a difficult area to ignore for the government of Bahia. Thus,
the state government created an agency called Alagados Melhoramentos S.A., (AMESA) with
" Note that the fact that the area was previously used a waste disposal and the importance that waste has had for the
stilts inhabitants, explain why the inhabitants of Alagados have a very special relationship with trash.
5
Interview with Sonia Fontes, CONDER president between 1994 and 1998.
the goal of implementing a major upgrading initiative in the area of Alagados between 1973 and
1983.16 The intervention, implemented by a group of committed and young public engineers and
architects - some of whom became also involved in the Ribeira Azul program - was consider
pioneer in Brazil because of its technical complexity and because it targeted a slum area.
Despite many delays for the high costs of the implementation (especially because of the land-fill
of the swampy area), and community's dissatisfaction with the size of the housing, AMESA
succeeded at eliminating all the houses on stilts in Alagados, built 2148 new houses, and 12
buildings between schools, health posts, and public offices. Additionally, it paved 300,000 m2
and filled with land a swampy area of 1,000,000 m2. However, as a result of the tolerance of the
governor Waldis Pires, and the extinction of AMESA, the whole area was reoccupied by the end
of the 1980s. Additionally, because of AMESA's control over invasions in Alagados, in the
beginning of the 1970s, the adjacent area along the bay in the Enseada dos Cabritos, was invaded
to absorb the high demographic pressures in the Peninsula of Itapagipe. For its similarity with
Alagados, this area became known as Novos Alagados (New Alagados).
Implementing Agencies: CONDER and AVSI
The Ribeira Azul project is implemented by the State Development Agency CONDER
and the Italian NGO AVSI. I will now present each of them in turn. CONDER was created by
ACM in 1974 to promote the infrastructural development of the Metropolitan area of Salvador,
and - after being under the person of the governor - was transferred under the Secretariat for
Planning and Technology (SEPLANTEC). In 1998, CONDER became the State Urban
Development Agency and incorporated the housing agency URBIS. In 2002, CONDER was
16 The
intervention was financed by the Banco National de Habitacao.
The team of engineers, architects and the social worker that worked for AMESA at the time, affirm that the
reoccupation of the area was probably facilitated by the empty land they left between the new houses and the shores.
"
transferred from the Secretariat of Planning and Technology to the Secretariat for Urban
Development (SEDUR). Since its foundation, as we have seen in section 1.2, CONDER has
played a very important role in the infrastructural development of the metropolitan area, and
more specifically in the construction of roads, highways, parks, as well as government buildings,
and the international airport. Because infrastructural development was the political priority of
ACM when he was in power, this agency has historically been controlled by his group and has
always been powerful and relatively autonomous. Additionally, because, in Salvador, the
Municipal government has historically been controlled by the opposition, CONDER represented
ACM's "municipal arm."
AVSI is one of the most important non-governmental organizations (NGO) in Italy, is
presently working in 35 countries of the world, is involved with more than 100 development
projects, and holds the General Consultative status with the UN Economic and Social Council
(ECOSOC) in New York and the UN Industrial Development Organization (UNIDO) in
Vienna.' 8 AVSI is part of two powerful economic and political institutions, the International
Catholic Movement "Comunione e Liberazione," and a non-profit group of Catholic
entrepreneurs "Compagnia delle Opere." Additionally, AVSI has a close relationship with the
Italian conservative party "Forza Italia."19
AVSI arrived in Salvador in 1989. At the time, AVSI was working with the Municipality
of Belo Horizonte, Brazil to upgrade an informal settlement of the city. Because of the success
18 The
consultative status allows NGOs to provide inputs to the ECOSOC into its deliberations at public meetings
and in its subsidiary bodies as well as in UN international conferences and their preparatory bodies. Each year the
approximately 2000 NGOs now holding consultative status receive the provisional agenda of ECOSOC. They have
certain privileges to place items on the agenda of ECOSOC and its subsidiary bodies; they may attend meetings,
where they may submit written statements and make oral presentations to governments
(http://www.ngocongo.org/ngopart/constat.htm).
19 The International Catholic Movement "Communion and Liberation" was supporting the coalition of the Prime
Minister Silvio Berlusconi, Forza Italia. (August 2004). Compagnia delle Opere represents the economic power of
"Communion and Liberation."
of the project in Belo Horizonte, AVSI was asked by the Archbishop of Salvador at the time,
Don Lucas, to contribute to the revitalization of the community of Novos Alagados through the
promotion of social activities. Such request of Don Lucas is explained by the close connection
between AVSI and the Catholic Church in Brazil. Don Lucas - who had at heart the situation of
the inhabitants of Alagados - was already supporting a number of activities in the area and had
close contact with the community organizations of the area. As a result, at its arrival in Salvador
AVSI entered the community of Novos Alagados thanks to Don Lucas' connections with the
community organizations of the area. The Community organizations of Novos Alagados and
Alagados have been particularly active since the military regime. This is the result of the effort
of the Brazilian Catholic Church to increase slum dwellers' political awareness, and of the
support that the Church granted to the mobilization of these community organizations to request
better water and sanitation services, drainage systems and roads (De Carvalho 2001). Currently,
there are about 70 community organizations in the area, mainly providing education and health
care to the inhabitants.
From the beginning, AVSI looked for a partner in the local government to start an
intervention in the area of Novos Alagados. Given that according to the Brazilian Constitution
the Municipality is responsible for the implementation of housing policies, AVSI initially
approached the Municipality. However, CONDER had greater interest and financial resources
than the Municipality, and in 1989, committed to AVSI to look for funding to develop a project
in Novos Alagados. At the same time, CONDER requested that the Municipality - at the time
governed by the opposition - would not to be involved.2 0 Thus, AVSI started working with
CONDER. The first project that saw the collaboration of AVSI and CONDER was a pilot project
in the Nova Esperanza area of Novos Alagados. This project, financed by the Italian Ministry of
20
Interview with AVSI and CONDER's official.
Foreign Affairs ($800,000) and by the State Government of Bahia ($800,000), benefited 15,000
people. The project, similarly to the project implemented in the 1980s by the government,
substituted stilt houses with houses on dry ground, provided water and sanitation infrastructure,
and the revitalization of the area along the shore.
The success of the pilot project in Nova Esperanza stimulated AVSI and CONDER to
look for additional funding to scale up the intervention to ultimately embrace the whole bay of
Novos Alagados. Through the Metropolitan Project, and then the Viver Melhor program,2 1 the
area of intervention expanded to Novos Alagados II etapa, Alagados I, II and Joanes CentroOeste. In order to better integrate and generate synergies among the different range of separate
initiatives in progress within the area, in 1998, CONDER and AVSI decided to scale up the
experience of Novos Alagados to the whole Ribeira Azul area.2 2 The different projects
implemented and ongoing in the whole Ribera Azul area became part of the Ribeira Azul
program. In 2000, the collaboration between AVSI and CONDER strengthened with the Italian
Government's approval of a $5 million grant to the State Government of Bahia to develop the
Social and Technical Assistant program (PATS). Not only the grant provided financial support
and technical assistance for the implementation of the Ribeira Azul program, but also establishes
a project management unit, made of CONDER and AVSI to ensure coherence among the
different interventions undertaken in the area (see the section about governance).
Financial Arrangement
The Ribeira Azul program is a $60 million program comprising a number of smaller
projects, financed by different national and international donors that invested in one or more
21 funded
22
through the Federal programs Habitar Brazil and Pro-moradia, started in 1996.
My interviews provided contradictory information about who and how proposed to scale up the program.
neighborhoods in Alagados and Novos Alagados. Figure 3 presents an overview of the amounts
and origin of the funds. 55% of the $60 million are provided in the form of loans by the World
Bank (30%) - through the Metropolitan Project and PRODUR project - and Caixa Economica
Federal (25%) - through the HABITAR/Brazil program. In addition, 21% is State Government's
funding, 12% is financed through Federal Transfers, 9% is a grant of the Italian Government
(PATS), and the remaining 3%is AVSI's private funding.
Figure 3: Financing of the Ribeira Azul Program (US$ million, 2004)
Grant of the
Italian
Gowmment
AVSI's priate
funding
$2
(3%)$1.
$5
ment of
Go
Bahia
(21%)
(9%)
World Bank
(PRODUR and
Projecto
Federal
Metropolitano)
Goemment
$16.9
(30%)
Transfers
$6.7
(12%)
Loan from IADB
through the
Caixa
-
Economica
Federal
$14.5
(25%)
The allocation of the funding among the different project components is shown in figure
4.
As expected, the majority of the funds go to infrastructure (65% of the $60 million).
Housing accounts 23% and technical assistance 2%. Social activities account 10% of the funds
($6 million), 40% ($2,4 million) and 34% ($2 million) of which are financed through the Italian
grant, and AVSI's private funds respectively. This suggests that the World Bank and Inter23
The calculation does not include the costs of the project inAlagados IV,since it was inprocess of financing,
or
the general costs of coordination, monitoring, supervision, and administrative management of PATS ($557,779). In
three neighborhoods figures did not distinguish between investments in housing and in infrastructure, I assumed that
investments in housing were one third of the total, and infrastructure the remaining two third. This assumption was
made considering that in the neighborhoods for which I had the data, infrastructure represented about two thirds of
the total costs of physical improvements. Furthermore, according to an interview with AVSI's management, the
investments in social activities are higher than 10%.
American Development Bank loans, and the funds coming from the State and Federal
government are sectoral24 in nature and include only limited funds - if any- to fund the social
component. 25
Figure 4: Allocation of the funding among different project components
(US$ 60 million, 2004)
Social activities
10%
Techical
assistance
2%
Infrastructure
65%
Housing
23%
Institutional Framework
I now move on to present the institutional framework and the governance structure of the
project. Figure 6 is divided into two parts, the upper one representing the institutions involved in
project financing and the lower one representing the institutions involved in implementation.
"Sectoral" is a technical term to indicate that loans are only financing activities in one sector, such as
2transportation or construction of roads, rather than financing infrastructure and social activites.
24
Financing through the Caixa Economica Federal always included funds to finance the social component.
Unfortunately, I did not have sufficient data to calculate whether and how much of the remaining 26% of the social
component is financed by whom.
Figure 6: Institutional Framework of the Ribeira Azul Project
Total = $60 millions
Note that while the money originating from the World Bank and the Federal government of Brazil go through the
State Government of Bahia, $12.2 million represents the contribution of the state to the project, and is not meant to
comprise the $16.9 and $6.7 million coming from the World Bank and the Federal Government of Brazil
respectively.
As shown by the green arrow, the World Bank (WB) loan goes through the Government of
Bahia. The red arrow shows that the Inter-American Development Bank (IADB) loan goes
through the Caixa Ec6nomica. The black arrow represents the Italian grant to the State
Government of Bahia, that is channeled through Cities Alliance (CA) to AVSI, who manages the
funds under the supervision of the WB.
26
Cities Alliance was launched by the World Bank and
United Nations Centre for Human Settlements (UN-Habitat) in 1999, and includes a number of
governments and multilateral institutions.
2
Its activities support the implementation of the
Habitat Agenda, and the goal of the Alliance is "making unprecedented improvements in the
living conditions of the urban poor by developing citywide and nationwide slum-upgrading
programs (Cities Alliance www.citiesalliance.org)." 2 8
The presence of three international institutions (WB, IADB, and CA), plus the Federal
Government, and the national Caixa Economica tend to make the project complex, in that each
of them has its own administrative and technical requirements. Furthermore, the complexity of
this financial arrangement is enhanced by the fact that the different funding agencies started
investing in Alagados and Novos Alagados at different times, and each focused on a different
area.29 As we can see in table 2, the World Bank financed two projects and started in 1986, the
IADB and the Italian Government one project each, and started in 1998 and 2001 respectively.
The State and the Federal Government also financed through the Caixa Economica Federal
(CEF) one project each.
Note that it is not common for a NGO to play such an important role and be able to comply with the World Bank
administrative requirements.
27 Cities Alliance include all the G-7 governments plus Brazil, Netherlands, Norway
and Sweden, the Asian
Development Bank, UN-Habitat, the United Nations Environment Programme, the World Bank, and local
authorities from different countries.
28 The Alliance aims at achieving the Millennium Development goal of improving the lives of at least 100 million
slum dwellers by 2020. Among the goals of Cities Alliance are: promoting projects that focus on the city and its
region rather than on sectors, supporting partnerships between local and national governments, and those
organizations directly representing the urban poor, and looking to scale up solutions promoted by local authorities
and the urban poor.
29 Such complexity, augmented by numerous changes of location of the CONDER offices since the project started,
may explain why it was strangely impossible to locate all the project documents or to have a precise picture of the
project outcomes. Consequently, a precise understanding of the amount of funding invested by each funding agency
in the different components, and within each neighborhood was not achieved.
26
Table 2: Ribeira Azul Program: donors, programs, years and areas
(in all cases the State Government of Bahia provided matc ing funds)
Area of destination
Year
Programs
Origin of the funds
Metropolitan
1986-1997
Aracas I e II, Novos Alagados I Etapa
PRODUR
1997-2004
Novos Alagados II Etapa, Alagados III
HABITAR/BID
1998-ongoing
Alagados IV, V, and V130, Mangueira
PRO-MORADIA
1995-ongoing
World Bank
IADB
(through CEF)
State Government of Bahia
Alagados I, II, Joanes Centro Oeste,
Joanes Azul, Mirantes
(through CEF)
Federal government
HABITAR/Brazil
1997-on going
Mudanza, Mangueira II
PATS
2001-2004
Ribeira Azul Area
(through CEF)
Italian Government
The lower half of figure 6 shows the program implementing agencies, which are the
Urban Development Agency of the State Government of Bahia, CONDER, the Italian NGO
AVSI, and within AVSI and CONDER, the Program Management Unit (PMU). Figure 7
summarizes the amount of funding that is managed by each agency. CONDER is responsible for
the allocation of $53 million, AVSI $2 million, and the Program Management Unit the remaining
$5 million. 3 1 Beside CONDER, AVSI and the Program Management Unit, the governance
structure of the Ribeira Azul program includes the Tripartite Committee and the Consultative
Committee. It is noteworthy to point out that the Municipality - legally in charge of housing
policies - is not included among the implementing agencies.
The project in Alagados IV will be financed by Habitar BID
The red arrow from CONDER and AVSI is explained by the fact that AVSI is subcontracted by CONDER for the
implementation of social activities in Alagados IV and V with funding coming from the Caixa Economica Federal,
as well as for the selection of personnel working in the Ribeira Azul program for CONDER.
30
3'
Figure 7: Funding managed by the implementing agencies and the Program Management
Unit (US$ million, 2004)
Total = $60 millions
Governance structure of the project
In order to better understand the governance structure of the program, it is worthy to take
a closer look at the Program Management Unit, the Tripartite Committee, and the Consultative
Committee. The PMU - established in 2001 by the World Bank and the Italian Government with
funding of the Italian grant-has a total of 13 workers and 2 co-directors, one for the
infrastructural component and the other one for the social component, each made of a
representative of AVSI and a representative of CONDER (see figure 8). Interestingly, the
decision making power is divided equally between CONDER and AVSI.
Figure 8: Program Management Unit
The PMU is responsible for the day to day management of the Social and Technical
Assistance Program (PATS), which comprises: technical and methodological coordination of the
entire program32 ; technical assistance for the development of the overall plan of intervention and
the necessary engineering and architecture plans; support to government operations in the area of
housing; execution of strategic social activities; systematization and publicizing of the
methodology; execution of studies toward the preparation of the citywide program Viver Melhor
II for the upgrading of informal settlements in the Salvador Metropolitan region (PATS
Implementation Plan 2001). The dashed line surrounding the Program Management Unit (PMU)
indicates that there is not a PMU office. Although the PMU is formally supposed to be located
within Conder's headquarters, and has a separate budget, AVSI and CONDER's workers - both
when they have PMU responsibilities or are directly working for the PMU - work in the
headquarters of their respective agency. The joint technical team of the PMU ensures quality of
Note that this is crucial giving the number of institutions involved, and the different
administrative and financial
requirements.
32
implementation, and consistency in methodology. This means that although AVSI only provides
a relatively limited percentage of financial resources to the project (3%, and 12% if we consider
the Italian grant), it has a very important role in the Ribeira Azul program. The additional
activities of the Ribeira Azul Program are executed under the responsibility of CONDER.
The work of the PMU is monitored by the Tripartite Committee and the Consultative
Committee. Both these institutions provide avenues for participation of different actors in the
project, including community representatives. The Tripartite Committee - formed by a
representative of the World Bank, a representative of the State Government of Bahia, and a
representative of the Italian Minister for Foreign Affairs (MAE) - meets every year to evaluate
and monitor the Technical and Social Assistance Project. The Consultative Committee is the
outmost participatory institution in the Ribeira Azul program, in that it comprises - beside the
members of the Tripartite Committee - five community representatives and the Municipality of
Salvador. Such structure aims at ensuring the participation and influence over the Municipality who is not among the implementing agency - and of community organizations. Community
representatives were chosen among the community leaders of Alagados and Novos Alagados and
were elected by community leaders of the 67 community organizations of the area. The
Consultative Committee meets every year to integrate community members and AVSI' concerns
into the program and to provide guidance for the PMU.
1.4 Conclusion
This chapter described the political context and urban development of the city of Salvador.
Additionally, it showed how the Ribeira Azul Program developed, the institutions involved, and
the history of the areas of Novos Alagados and Alagados. Additionally, it presented the complex
financial and institutional arrangement, and the outcomes of the project. Finally, it provided an
overview of the participative governance structure of the project, which is designed to ensure the
involvement of community organizations and the Municipality in project implementation.
Chapter 2
Top-down approach going hand-in-hand with construction
2.1 Introduction
Slum upgrading calls for a mix of physical design, putting definitions to public and
private spaces, and in some cases, relocation. Finding this mix stands to benefit from community
input. The task of convincing residents to resettle, or even "only" cede portions of their lots to
allow for road widening or a similar project is difficult and is said to be possible only through
beneficiaries' participation in project design and decision making (Nance 2004 and Imparato and
Ruster 2003). Since NGOs are said to have a comparative advantage in understanding and
responding to local needs, they often serve as intermediaries between the centralized
infrastructure agency implementing the slum upgrading projects and the community, helping to
plan and organize the slum upgrading intervention (Clark 1991, Nance 2004, Imparato and
Ruster 2003).
The World Bank, Cities Alliance, and the Italian Government believe that community
participation was one of the most important reasons for the success of the Ribeira Azul program.
More specifically, they argue that AVSI played an essential liaison role between the community
and CONDER, ensuring community participation in the project. However, I argue that in spite
of the project being considered a success, the assertion than it was participatory lacks support.
Section 2.2 provides evidence as to the lack of community participation in project design.
Section 2.3 discusses how CONDER and AVSI succeeded at building infrastructure and houses,
in spite of the lack of community participation in project design. Section 2.4 presents two
different institutional arrangements - one involving the state agency CONDER and the other one
involving both CONDER's engineers and AVSI's social workers - that at different points in time
successfully provided individualized technical solutions to accelerate the construction of
infrastructure and housing.
2.2 Limited communityparticipationin project design and decision making
While the Ribeira Azul project is considered by the World Bank, Cities Alliance and the
Italian government as an exemplary example of a participatory project, the complaints of
community organizations' and AVSI's field staff about the concept of participation adopted by
the project, and lack of community members' involvement in project design, suggest that the
project is more "top-down" than participatory.
Differing Views of Participation
Community members lament about the way AVSI and CONDER conceive participation.
In a letter written by the community organizations of Alagados and Novos Alagados to the
Consultative Committee, 33 community members requested that the Committee clarify the way
the program interprets the concept of "participation" given that in its application they "do not
find any correspondence with the concept of participation" they know. 3 4 Note that a similar
complaint was exposed by the community organizations of Novos Alagados and Alagados in a
previous open letter to the City of Salvador. According to this letter, the community considers
participation as "a process in which the community expresses its opinion and its decision making
right." More specifically, the letter refers to the projects of Alagados IV and Mangueira as
examples of projects that lacked opportunities for community discussion in the planning phase,
The Consultative Committee is formed by high level representatives from the World Bank, the Italian
Government, the Governor of Bahia, 5 community representatives, AVSI, and CONDER.
3 This letter is dated 23 March, 2004 and was presented during the Consultative Committee in the presence of the
World Bank, the Italian government, AVSI, and CONDER.
33
and about which "nobody [among the inhabitants] knows anything." The letter also stated that
"community members and organizations' criticisms of the project were not taken into account",
or are "taken personally" and that "associations have been invited as token participants and not to
influence decision making, since there is no real dialogue." Interestingly, this complaint is
similar to what was expressed by more than one field staff. Project staff describe their meetings
with the community as "informative meetings, in that the community is informed about decisions
that have already been made by CONDER and AVSI's high level officials." Additionally, they
affirm, "AVSI's concept of participation is to have the community agree, not to participate."
These complaints suggest that community members do not feel they have a stake in the decision
making process, and that their complaints are validated by some members of the project staff.
Furthermore, in a letter written to the Italian Minister for Foreign Affairs 35 , the
Consultative committee, and the city of Salvador, community members suggest that they lack
information about the Ribeira Azul program. More specifically, in the letter to the Italian
Minister for Foreign Affairs, the 67 community organizations of Novos Alagados, and of the
network Comissao de Articulacao e Mobilizacao dos Moradores da Peninsula de Itapagipe
(CAMMPI) requested access to some documents of the Technical and Social Assistance Program
(PATS) in order to discuss the way the Ribeira Azul project has been implemented.36 The
community organizations' request for more information is not surprising, given that within the
Consultative Committee the role of the five community representatives is to review
implementation progress, and provide suggestions and overall guidance to the Program
Management Unit (PATS Implementation Plan 2001). Precisely in order to perform their
This letter is dated 18 September, 2004
Specifically, community organizations requested access to the following documents: the PATS project document,
the budget, origin, and allocation of the PATS project funding, and the list of the guidelines of Cities Alliance that
inspired the design of the project.
3
36
assigned role, community representatives again requested in a subsequent letter to the
Consultative Committee information about the activities already implemented, populations that
have benefited, and activities to be implemented by the project. Furthermore, community leaders
wanted to see these documents in order to ensure that the Italian grant was used, as promised by
the project staff, to finance the construction and operation of schools and community centers
owned and run by the community. 37
Community Members' Lack Influence on Housing Design
Community members' limited influence on the design of the housing represents another
way in which the project is more "top down" than international donors believe. Community
members and leaders assert that CONDER and AVSI are "indifferent" towards their housing
design concerns. Community members express dissatisfaction with housing design for five
reasons. First, they see the house size offered by the project as very limited, often smaller than
the stilt houses where they live. The size of the houses built varies, depending on the area and
the space available, between 14 m2 and 48m 2 (specific increments include 16 m2, 20m2, and
22m 2 ), but even the larger houses offered by the project are very small for a large family.
Further, it is not automatic that the project provides larger families with larger houses. Second,
the limited lot size often hinders families from expanding their houses, if needed. Third, the
internal layout, comprised of one large room, limits privacy, which is particularly problematic
for large families. Additionally, many community members dislike having a bathroom located by
the house entrance. Fourth, houses do not always have independent walls. Community members
Community organizations wanted to ensure that the Italian grant was not used to finance schools and institutions
owned by AVSI, instead of financing organizations owned by the community. Note that AVSI's manager affirms
that AVSI's schools and community centers have been built with AVSI's private funding. This evidence derives
from interviews with community leaders and the letter to the Consultative Committee of the Ribeira Azul project
and the Bahian Authorities dated March 23, 2004, and signed by the 5 community representatives accredited in the
PATS Consultative Committee.
37
value independent walls for their ability to buffer their homes from the surrounding noisy
environment. Finally, some of the project houses have two-stories, which community members
see as unfavorable, as this layout blocks their ability to build vertically to accommodate growth
in family size. Note that building vertically when the family enlarges is a cultural habit of poor
families living in the area.
Both AVSI and CONDER high level officials gave limited importance to these
comments, on the grounds that the housing design offered by the project is the only possible one.
For example, AVSI's manager believes that the limited range of housing options and the
technical difficulty of the project leave little room for community participation in project design.
Additionally, he claims that the requests of the community to modify the design of the houses are
financially and technically unfeasible. "Community members only think about themselves" and
"do not seem to realize that if we build larger houses, we are going to be able to build less of
them, and that some people will have to be resettled."
Given the unavailability of land in Alagados and Novos Alagados, and CONDER's
effort to minimize resettlement, I believe that there is indeed a trade off between building bigger
houses and resettling inhabitants out of the area. At the same time, I believe that a greater
consideration for community members' concerns as least in so far the location of the bathroom
and the independent walls, as well as more constructive dialogue about the reasons why a
different housing design would not have been possible, could have made a difference in
community members' attitude toward the project.
Why does AVSI Play only a Limited Intermediary Role
Although the lack of trust toward AVSI may to some extent derive from community
organizations' different political orientation, or unmet expectations of playing a greater role in
the design of the project, the complaints of community organizations do show that the project has
not been as participatory as believed. Particularly since AVSI gained recognition as a gobetween, it is surprising that AVSI did not play an intermediary role between the community and
CONDER. Indeed, in contrast to the common assertion that NGOs are "participatory," this does
not seem true here, at least in terms of community members' participation in project design and
decision making. I argue that AVSI's limited intermediary role has three possible explanations:
(1) community organizations' perception of AVSI as lacking independence from CONDER, and
the government more in general; (2) limited interest of AVSI to advocate for the community, for
threat of compromising its relationship with CONDER, and (3) a seeming tradeoff between
community participation and scale of the project.
First, community leaders and members do not see AVSI as a separate organization from
CONDER and perceive the boundaries between CONDER and AVSI as blurred. For example,
one community leader said: "CONDER and AVSI are the same thing, they think in the same
way, they are not separate organizations." Another community leader said, "AVSI lost its
identity as an organization close to the community, and that's bad." This community leader went
further explaining that AVSI acts as if it is an "arm of the government." However, another
interviewee affirmed that a difference exists between AVSI and CONDER in terms of the values
and mission of the two organizations. But, he said, "even if AVSI listens, it does not act on the
basis of what it listens to." These comments suggest that AVSI is not particularly successful at
responding to community members' needs and that NGOs are not always "closer" to people than
government agencies are. Furthermore, the community perceived that only one organization,
CONDER, implemented the project. For example, community members refer to AVSI's social
staff in the field as "girls of CONDER."38 Additionally, all of the letters of complaint written by
the community are addressed to the project staff, and not directly to AVSI. This holds true even
when complaints are specifically directed to AVSI.
Second, I believe that because AVSI gains an important role in project implementation on
account of a strong relationship with CONDER it has a limited interest in advocating for the
community if that threatens its relationship with CONDER. For example, CONDER
subcontracted AVSI to develop social activities and social intermediation in Algados IV and V
and AVSI is also the current co-director of the project. Maintaining a close relationship with
AVSI is in the interest of AVSI to survive and grow as an organization. According to high level
CONDER's officials, CONDER would not have allowed AVSI to have such an important role in
project implementation, if "it did not felt AVSI was a loyal ally, closer to CONDER rather than
to the community."
Third, my interviews suggest a possible trade-off between participation level and scale of
the project. According to an AVSI staff member who participated in the Novos Alagados project
before it scaled up to the entire Ribeira Azul area, "it is impossible for AVSI to maintain the
same close relationship with the community that it had in Novos Alagados when the area of the
project and the number of activities implemented were much smaller." CONDER's social
director of the Ribeira Azul project, who also participated as a member of AVSI's staff in the
project in Novos Alagados, again suggested that "the large number of donors, the different
administrative requirements they have, and the large number of activities implemented by the
project do not allow the same relationship with the community that AVSI had when the project
started." Along these lines, AVSI's management affirmed that, because of the project size, they
now need to work with community associations rather than with community members. Note
38
From the Portuguese "Meninas da CONDER"
however, that this represents a problem if the issues discussed with community leaders are
affecting individual properties. For example, one of the community leaders of Alagados
suggested that "the project staff should not discuss the design of the houses with community
leaders but directly with community members to ensure that houses meet community members'
needs." Thus, while the increase of project size results in the need to work with community
leaders rather than with single community members, the feasibility of this approach seems
limited when the issues discussed affect individual goods such as housing.
2.3 BuildingInfrastructure and Housing Without Community Participation
in Housing Design
While community participation is considered necessary for the implementation of slum
upgrading projects, as we have seen, community participation was limited in the Ribeira Azul
project, at least in respect to project design. So, how do we explain that the project was able to
eliminate 1,700 stilt houses, build a costal road, connect the settlement to the main road network,
and provide 7,700 families with improved water and sanitation infrastructure? I argue that
CONDER and AVSI succeed at building housing and infrastructure for three reasons: 1) the
living conditions of the inhabitants of Novos Alagados and Alagados, 2) project design, and 3)
the manner in which the project was implemented.
The Living Conditions of Alagados and Novos Alagados Inhabitants
The living arrangements of the inhabitants of Novos Alagados and Alagados prior to the
launch of the project facilitated project execution for a number of reasons. First, many people
living in the stilt houses survive in inhumane living conditions (see picture Appendix II): houses
on stilts are made of pasteboard, lie in and are surrounded by trash, lack stable foundations, have
broken roofs, and often lack safe water access and sanitation facilities (see appendix III).
Additionally, stilts needs to be substituted every 4-6 months because the tides rot the wood, and
to further exacerbate the difficult situation, wood has become increasingly expensive over time.
Moreover, replacing stilts requires male laborers, but many households in Novos Alagados and
Alagados are female headed. Therefore, the project presented these women with a unique
opportunity to both move into a house and solve many of these other related problems. Finally,
because stilt houses built around the same pier structurally support one-another, when one or
more are removed, the others are likely to fall. Thus, when one or more families of a pier
participated in the project, it resulted in the participation of the entire pier's inhabitants (see
appendix III).
The role of Project Design
The design of the project helped the construction of housing and infrastructure in two
ways. First, CONDER decided to minimize 39 the number of families to resettle to avoid the
difficulties typically associated with resettlement, by land-filling the tidal area where the houses
on stilts were built. This occurred despite the high cost of land-filling (Teixera 2002). For
example, as contrasting pictures of Novos Alagados in Appendix II demonstrate, the area where
stilt houses were present in 1984 used to be tidal, but was then filled during the project.
Resettlement is often one of the most challenging components of slum upgrading projects
because it requires intense negotiations with the community. Slum dwellers regularly refuse to
relocate or leave the land that they have occupied for many years, often illegally. This was
especially true in the case of Alagados and Novos Alagados; the extremely poor living
conditions of its inhabitants attracted great attention and became well-known across the country.
Moreover, because many poets and artists had celebrated the charm and uniqueness of these
39
Indeed, the number of families resettled is only 136 out of a total of 800 that were substituted their house on stilts.
slums, they had not only contributed to their fame, but also created a feeling of pride and
attachment to the area among the inhabitants. Thus, by avoiding resettlement CONDER and
AVSI succeeded in having the community to support the project. Second, CONDER indirectly
pressured the inhabitants to buy in the project because of its decision to stop paying to maintain
the piers that held the stilt houses together (see Appendix III). Note that costs of maintenance of
piers are high both in monetary and social terms because the inhabitants need to organize to buy
the material, come together to build or repair them, and share the financial burden.
The role of Project Implementation
Characteristics of project implementation facilitated physical improvements in two ways.
First, while project beneficiaries are formally supposed to pay for the houses, CONDER,which is
formally responsible for payment collection, has never collected the payments.
While this is
typical of many slum upgrading projects, it played an important role in facilitating the
implementation of the project in that many community members said that they would have never
accepted such a house if they had had to pay for it. Interestingly, community members declare
that they would never pay for the house, even at the moment they sign the contract in which they
agree to have their house substituted and commit to pay for the house. According to CONDER's
staff, CONDER does not have the institutional capacity to collect payment. There are, however,
other more important reasons that explain why CONDER never collected payments: lack of
political convenience and high costs of collection of the payments. CONDER finds it politically
convenient not to have inhabitants pay for the houses. Indeed, CONDER has used the project for
political purposes on more than one occasion, the most important of which took place in 1997,
Households are supposed to pay a monthly rate of 10% of the minimum salary for 5 years and they commit to pay
these monthly rates when they sign the contract in which they take responsibility for paying for the house.
40
when Sonia Fontes, the then president of CONDER, left to become state depute for the PFL4
party. Under such circumstances, in order to ensure Fontes the votes of the inhabitants of Novos
Alagados, CONDER forgave the payment for the houses built by the project until that year. The
other reason why CONDER does not collect payments is that - because of the very low prices
that project beneficiaries pay for the houses - it would be more expensive to collect the
payments, rather than forgive them. Indeed, in order to collect the payments, it is necessary to
have field agents that go door to door every month, and administrative support to keep track of
each inhabitant's payment record.
Second, when implementing infrastructure and housing improvements, engineers made
small changes in project design to accommodate community members' needs. The physical
design of these changes ranged and included relocating the door or a house in response to a
request. Another case involved a slight narrowing of a coastal road that the project is building in
Alagados III, in order to reduce the impacts it would present in passing over the plot of an
unemployed woman living with 8 children in a stilt house. The project staff weighed the
circumstances of this woman and decided to avoid forcing her to resettle. Note that project staff
considered these modifications a way to bridge the "distance" between the community and the
project, and compensate for community members' discontent with the lack of involvement in the
design of the housing. For example, one social worker said, "we can't do much because high
level officials won't listen to us, but at least we do as much as we can to help these poor people,
and make the project to happen." This social worker's comment illustrates social intermediation
specialists' frustration with the limited influence of community members in the housing design.
Furthermore, this comment suggests that social intermediation specialist see themselves as
having the crucial role of making project implementation possible.
41
PFL stands for Party of the Liberal Front.
2.4 Community participationin project implementation: two different
successfulinstitutionalarrangements
As we have seen, getting residents to agree to cede parts of their lots to widen a road, or
build a drainage system is difficult and often requires relaxing technical standards and finding
individualized solutions. It is therefore worth taking a closer look at the two different
institutional arrangements used in the Ribeira Azul program that succeeded at providing
individualized solutions to meet community members' needs and accelerated project
implementation.4 2 The first case characterized the implementation of the project from 1993 to
2000 in Novos Alagados and represents an example of a centralized institutional arrangement. In
this example, CONDER's socially-minded technicians and a social intermediation professional
work in close contact with the community.43 The second case characterizes the implementation
of the project from 2000 on in the areas of Alagados III, IV and V and illustrates a decentralized
model. In this latter example, AVSI's social intermediation workers serve as liaisons between
CONDER's technicians and the community. The move from a centralized to a decentralized
institutional arrangement in 2000 has two motivations. First, in 2000, the leader of CONDER's
technical team working in the Ribeira Azul project resigned as a result of tensions with
CONDER's new president, Mario Gordilho. Subsequently, Gordilho dismantled CONDER's
team and substituted it with engineers that had little, if any, experience working in slums. He
also increased the role of subcontractors as social intermediation specialists. Second, around
Note that, as mentioned in the previous section, these modifications are small and do not include modifications of
the housing design.
43 Note that AVSI was present in the area even before 2000, but the division of labor between AVSI and CONDER
was such that CONDER's technicians and social worker were also deeply involved in social intermediation
activities.
42
2000, AVSI assumed exclusive responsibility for social intermediation in the areas of Alagados
III, IV and V, thus becoming the intermediary between CONDER and the community.44
The first of the following sections will analyze the "centralized" institutional
arrangement. The second one will briefly present the events that occurred in 2000 and discuss
the "decentralized" institutional arrangement.
1993-2000: "Centralized" provision of customized solutions
Large, centralized government agencies often have difficulty working in poor settlements
due to high population density, unplanned housing layouts, and a lack of respect for urban design
standards. These challenges all give rise to a need for individualized solutions that meet the
needs of community members. Despite being a large and centralized agency, CONDER
succeeded in providing individualized and flexible solutions according to community members'
requests. I argue that CONDER's success is the result of the presence of a group of socially
minded and committed engineers and architects, as well as a social intermediation professional,
who benefited from autonomy and discretion granted to them by CONDER's president Sonia
Fontes.
When the project started in Novos Alagados in 1993, the then current president of
CONDER, Sonia Fontes, asked the engineer Medici to take responsibility for the infrastructural
and housing improvements of Novos Alagados. Medici was the central figure responsible for the
design and implementation of the first government project in Alagados, executed in the 1980s by
the state agency AMESA, and therefore had significant experience in working in Alagados. 45 As
44AVSI was contracted by CONDER for the implementation of Alagados IV and V. The project in these areas is
financed by the Caixa Economica Federal. Instead, in Alagados III, social intermediation is paid through the grant
of the Italian government.
45 As stated in chapter 1, AMESA stands for Alagados Melhoramentos, and it was a state agency created with the
only goal of building houses and infrastructure in Alagados.
stated in chapter 1, the state government project implemented by AMESA succeeded in
removing all of the stilt houses in Alagados. It also gained notoriety in the country for the
techniques used to land-fill the swampy regions, and for being one of the first slum upgrading
projects in Brazil which succeeded in limiting resettlement, and providing water and sanitation
infrastructure to slum dwellers. At the time, the AMESA office was located in Alagados,
surrounded by houses on stilts. 46 Medici and one of the other members of his team recall that,
when working with AMESA, they spent days and nights talking with community members,
trying to meet community members' requests of small modification to project design so that the
project could replace their stilt houses with homes on dry ground, and solve conflicts among
community members.
Because of his reputation and experience, Sonia Fontes granted Medici unwavering
support as well as autonomy to organize and implement the project. For instance, Fontes
followed through on two requests made by Medici: establishing a small CONDER's field office
in Novos Alagados where Medici and his team would have worked, and allowing him to choose
his own team members. Medici selected four people to work with him: another engineer, two
architects, one of which had worked with him in the 1980s in Alagados, and one social worker.
From the beginning, all of the team members were located in the field office and had direct
contact with community members on a daily basis. Such an operational structure was unusual
for CONDER's workers, especially for engineers and architects. All of the members of the team,
even Medici himself, recall with enthusiasm and positive sentiment that they spent most of their
days talking to community members about the project, and finding technical solutions that could
satisfy the requests of these people on the ground. Indeed, some members of the team told me
Note that this office was very important for the population because it represented the presence of the state at the
local level. Community members went to this office for a variety of problems.
46
that what distinguished them from CONDER's high level officials and technicians was their
"social culture." Even those members of the CONDER group that did not participate in the first
Alagados project in the 1980s, developed technical expertise to work in such an environment,
and grew attached to the place. They likewise showed commitment to their jobs. I was struck by
these workers' relationship with the place and the communities of Alagados and Novos
Alagados, and by their sense of ownership of the project, which they referred to as "their baby."
They described working in Alagados as "much more than a job," and added, "we were married
with Alagados." As illustration of the high degree of commitment, some described how even
during the rainy season, these public officials often spent their lunch breaks and weekends
checking to ensure that the drainage system they had designed was working properly.
Additionally, they often resolved conflicts among community members, and even took
community members to the hospital on more than one occasion. The findings presented in this
section suggest that the presence of a socially-minded group of engineers allowed a centralized
agency, such as CONDER, to provide individualized solutions and accommodate community
members' needs.
2000 to present: "Decentralized" provision of customized solutions
The second institutional arrangement used in the Ribeira Azul project to provide small
modifications to project design to accommodate community members' needs is characterized by
the presence of AVSI's social intermediation workers as liaisons between CONDER's
technicians and the community. This institutional arrangement relates to project implementation
from 2000 on in the areas of Alagados III, IV and V. Although NGOs are said to operate
flexibly and in a decentralized fashion, and are therefore suited to play an intermediary role
The design of the drainage system represents a particular challenge in the area from the technical and engineering
view point.
47
between infrastructure agencies and community members, close collaboration between
technicians, such as engineers and architects, and social intermediation professionals working for
NGOs is typically a challenge (Davis et. al 2001). Engineers often refuse to work with social
intermediation professionals because they consider this group of a low professional status and
lacking knowledge about technical issues. On the other hand, social intermediation specialists
often express frustration at not being listened when they present technical suggestions to
engineers. In contrast to what normally happens, in Alagados III, IV, and V, AVSI's social
intermediation specialists work successfully with CONDER's engineers and architects to provide
individualized solutions that satisfy community members' needs.
In order to understand how AVSI's social intermediation specialists and CONDER's
technical staff successfully worked together, it is necessary to provide an historical backdrop to
the events that occurred in the Ribeira Azul project in 2000.
In 2000, AVSI assumed exclusive
responsibility for social intermediation in the areas of Alagados III, IV, and V. 48 This partially
arose as a result of a decision by Gordilho, the new CONDER's president, to subcontract AVSI
for the implementation of social activities and social intermediation in Alagados IV and V. It
also stemmed from the fact that social intermediation specialists in Alagados III, are paid
through the Italian grant managed by AVSI, and are therefore AVSI's employees. Second,
Medici resigned 49 as a result of his tensions with the new CONDER's president Mario Gordilho,
AVSI was contracted by CONDER for the implementation of Alagados IV and V. The project in these areas is
financed by the Caixa Economica Federal. Instead, in Alagados III, social intermediation is financed by a grant
from the Italian government.
49 Medici's resignation was the result of three events. First, Gordilho decided to close CONDER's field office in
Novos Alagados and transfer the project unit from Novos Alagados to CONDER's headquarters. This decision was
motivated by Gordilho's desire to maintain control over the project and over Medici's team. The team was well
known in CONDER and in Bahia for its technical capacity and commitment, but it was also left-wing, had a
relationship with all community organizations in Alagados and Novos Alagados, and was accustomed to having
autonomy and discretion. According to Medici, Gordilho's decision "killed the soul of the project" because "it did
not make any sense to work on the project from the headquarters," and with an unmotivated the team. Additionally,
the team took Gordilho's decision as a demonstration of lack of trust and lack of understanding of the importance of
48
and Gordilho, after dismantling his team,50 replaced it with technicians that had little, if any,
experience working in slums. Indeed, the Ribeira Azul project was the only slum upgrading
project implemented by CONDER until recently, and most of CONDER's technical staff has
experience only in building infrastructure, such as streets, waste disposal plants, and houses on
empty land.
I argue that CONDER's engineers and AVSI's social intermediation specialists
successfully worked together for two reasons. First, CONDER's engineers realized that
coordination with AVSI's social staff made their work easier. In fact, as explained in previous
sections, the newly appointed CONDER's engineers lacked experience working in slums, and
were not accustomed to close interactions with community members. Some even considered it a
"punishment" to work in the field offices in Alagados and Novos Alagados. Second, CONDER
engineers and AVSI social workers informally interacted on a daily base during lunch breaks.
The project pays for lunch for the staff of both institutions in a commercial center in the area.
This commercial center is the only place where workers can eat in the area. Such lunch breaks
have stimulated dialogue between engineers and social staff of the two agencies, and have
facilitated understanding of each other's way of thinking. Many engineers told me that these
conversations and daily interactions with AVSI's staff and community members has changed the
their work in the community. Second, Gordilho's acceptance of the proposal of the Italian Government, the World
Bank and AVSI to channel the $5 million grant offered by Italian Government to AVSI, rather than to CONDER.
Indeed, Medici wanted to ensure that CONDER maintained control over project funding and was concerned about
AVSI's increasing role in the project. Third, Gordilho's disregard of Medici's denounciation that one of the
construction firms doing the land-fill in Novos Alagados had been charging CONDER higher costs than what it had
defrayed. Despite the fact that Medici proved that the construction firm had been charging between 10% and 15%
more than the cost of the land-fill, Gordilho did not support Medici. This is because Gordilho and CONDER have a
close relationship with the construction sector, and more specifically with that firm.
50 One engineer is still working in the Ribeira Azul project in CONDER's headquarters. The sociologist is now
working for CONDER in the interior of Bahia. An architect resigned and now works at the University of Bahia. The
other architect is currently working as an advisor to the Minister of Cities in Brasilia. Finally, Medici is doing a
PhD at the Escola Politecnica da Bahia, and is involved in the preparation of the next government project Viver
Melhor II in CONDER's office. However, both Medici and the other engineer still working on the project have
limited autonomy and responsibility compared to the past.
way they see their jobs and it has changed them as people. In conclusion, a "decentralized"
institutional arrangement is also successful at ensuring that small modifications of project design
satisfy community members' needs. Specifically, such decentralized arrangements seemed to
have worked on account of engineers' need to cooperate with social intermediation professionals
in order to simplify their job.
2.5 Conclusion
This chapter looked closely at community participation, considered one of the primary
reasons for the success of the Ribeira Azul project. It not only provided evidence that the project
is less participatory than what the World Bank and Cities Alliance indicate, but also that AVSI
was not successful at ensuring that community members' concerns were factored into project
design. This chapter also provided evidence that infrastructure and housing improvements can
be implemented without community participation in project design. More specifically, it showed
that CONDER and AVSI succeeded at building infrastructure and housing because of the
following: (1) the extremely difficult living conditions of the inhabitants of houses on stilts in
Alagados and Novos Alagados; (2) the limited number of inhabitants resettled; (3) CONDER's
indirect pressure on the inhabitants to participate in the project by ending their financial support
for pier maintenance; (4) CONDER's decision to not collect payments for houses; and (5) the
implementation of small modifications to project design in order to accommodate community
members' requests. Furthermore, this chapter showed that both a "centralized" and a
"decentralized" institutional arrangement can be successful at providing customized solutions to
community members' needs. Finally, the chapter suggests that in order to draw lessons for
policies, community participation must be deconstructed into the different forms it takes.
Chapter 3
The Politics of Maintenance
3.1 Introduction
Maintenance typically represents a challenge in many infrastructure, including slum
upgrading projects. This is also true in the case of Ribeira Azul. This chapter aims at explaining
the reasons why maintenance has been a problem in the Ribeira Azul project. Section 3.2
presents the reasons encountered in the literature about why maintenance represents a problem in
many projects, and describes the specific situation of the Ribeira Azul project. Section 3.3 and
3.4 discuss the political factors - namely inter-governmental competition between the State and
the Municipality, and party politics - that explain why the Municipality is not performing its
maintenance role. Section 3.5 presents the positive aspects associated with the lack of
involvement of the Municipality. Section 3.6 concludes.
3.2 The exclusionof the Municipality
While maintenance of infrastructure is crucial to ensure project sustainability of the
investments made and the networks and equipment installed, it typically represents a challenge
for many reasons. First, agencies find maintenance politically less attractive than construction
(Tendler 1993). Second, engineers often consider maintenance a diminution in terms of
standards, prestige and past work and therefore often refuse to perform, or badly perform
maintenance tasks (Tendler 1993). Third, projects that involve mainly construction often do not
incorporate guidelines for project maintenance do not allocate funding or provide technical
assistance, or administrative support to the institutions responsible for maintenance (Bamberger
et al. 1982). Furthermore, international donors, such as the World Bank, do not provide
incentives for maintaining already built infrastructure projects in that they do not tie successive
disbursements of loans or grants to the maintenance of what was previously built. Note that
maintenance can represent even a greater challenge when it is performed by a separate agency
than the agency responsible for construction. When the institution responsible for maintenance is
separate from the institution responsible for construction, the two agencies need to cooperate.
However, coordination often represents a significant challenge, especially if the institution
responsible for maintenance has not had a stake in project design and implementation or if it has
not received funding and equipment from the project to perform its maintenance tasks.
Responsibility for maintenance of infrastructure built by the Ribeira Azul project stays
with the Municipality of Salvador. Problems related to the poor maintenance of infrastructure
and conflicts among CONDER and the Municipality for the responsibility of maintenance have
already manifested and put the sustainability of the project at risk. For example, in Novos
Alagados, CONDER and the Municipality are both neglecting their responsibility to fix a main
road that, when invaded by rain water, cannot be used by residents because it practically
resembles "a river". According to CONDER, the Municipality is responsible to fix the road,
because the Municipality is responsible for maintenance. On the other hand, the Municipality
affirms that the responsibility for the repair of the road stays with CONDER because the road is
inundated by water as a result of a mistake made by CONDER when the road was build.
Lack of maintenance in the Ribeira Azul project can be strongly attributed to the fact that,
despite being the institution legally responsible for maintenance, the Municipality of Salvador is
de facto not included among the implementing agencies of the Ribeira Azul program, nor seems
to have any stake in project design or implementation.51 Formally part of the Consultative
committee, the Municipal representative has changed numerous times, and while I was in
Salvador I was not able to identify him. Even the Italian Grant, which aims - among its goals at providing technical assistance and building institutional capacity in the local government, did
not provide any technical assistance, in the form of equipment, training courses, personnel, or
simply funding to the Municipality. The exclusion of the Municipality from the implementation
of a slum upgrading project is even more surprising if we consider that under the 1988
Constitution, Brazilian municipalities gained exclusive authority to organize and provide local
public services, to organize municipal zoning laws, and to legislate and develop urban
development plans. Additionally, the Estatuto da Cidade, Law 10.257/01 further devolves the
responsibility of developing local urban development policy solely to the municipalities.
The World Bank has repetitively tried to convince politicians and CONDER's officials to
provide a greater role for the Municipality. For example, the Task manager of the PRODUR
and of the Ribeira Azul project had private conversations with CONDER's high level officials,
Antonio Carlos Magalies (ACM) and the other Party of the Liberal Front (PFL) governors Paolo
Souto (PFL, governor between 1995-1999 and between 2002-2006), and Borges (PFL, governor
between 1999-2002). However, the World Bank has not succeeded at increasing the
involvement of the Municipality in project implementation. Furthermore, the Municipal
representative in Consultative Committee has never been an influential Municipal public official,
and as a result collaboration between CONDER and AVSI and the Municipality has been very
difficult. However, despite the request of the World Bank to appoint an influential officer within
the Municipality as the person responsible for the implementation of the Municipal tasks related
to the Ribeira Azul project - nothing has changed as of August 2004.
5
From now on, the Municipality of Salvador will be simply called "Municipality."
So, how do we explain the exclusion of the Municipality from the implementing agencies
of the Ribeira Azul project and the limited collaboration between the CONDER and the
Municipality? I argue that such exclusion stems mainly from the inter-governmental political
competition between CONDER and the Municipality and party politics. I will now discuss each
of them in turn.
3.3 Inter-governmental politicalcompetition
Inter-governmental political competition between the State and Municipal government
led to the exclusion of the Municipality from project implementation since in the initial
negotiations for slum upgrading in Novos Alagados in the early 1990s. At the time, CONDER
was responsible for the development of the entire metropolitan region of Salvador, and was
implementing the Metropolitan Project - a $77 million project financed by the World Bank and
aiming at improving infrastructure, waste management systems, solid waste collection, and
sanitation facilities in a number of areas of the city of Salvador. CONDER started its
intervention in Novos Alagados in 1993, after ACM decided to include Novos Alagados as one
of the areas of the Metropolitan project. In 1993, the Municipality of Salvador was governed by
Lidice Da Mata (1993-1996), a member of the PSDB, the opposition party, whereas the State
government was governed by ACM (1991-1995). Because the Municipality of Salvador was
governed by the opposition, ACM's decision to exclude the Municipality from taking part to the
implementation of the Metropolitan project, and to the project in Novos Alagados is not
particularly surprising. In fact, at the time, ACM and the PFL party precisely pursued the
strategy of limiting financial transfers from the State to the Municipal level in order to impede
Lidice Da Mata's governance of the Municipality. Salvador has the poorest record among all
Brazilian state capitals Brazil in terms of revenues collected,s2 and historically mayors have had
to rely heavily on inter-governmental transfers, or on financing from other groups such as the
construction industry, to implement their policies (Souza 1997, 1999). Political competition
between the State and the Municipality, ACM's political influence in Salvador and Bahia, and
the high dependency of the Municipality on inter-governmental transfers brought about the
failure of Lidice Da Mata's mayoral administration. The impact of the PFL's strategy is evident
from the fact that with the election of the PFL candidate Imbassahy in 1996, just a year later, the
Municipal revenues from transfers increased 70,8% (Fernandez 2004).
As soon as he was elected, Mayor Imbassahy (1996-2004) declared his loyalty to ACM,
and his willingness to establish a close relationship with Governor Paulo Souto (1995-1999), also
from the PFL party (Fernandez 2004). Yet the Municipality remained excluded from the Ribeira
Azul project. How do we explain the exclusion of the Municipality from the Ribeira Azul
project, given the PFL control of both the State and the Municipality and Mayor Imbassahy's
interest in a closer collaboration with the State Government? Party politics is the other main
reasons of the exclusion of the Muncipality from the Ribeira Azul Project.
3.4 Party politics
Party politics, and more specifically important political decisions of ACM - the leader of
the PFL party - and the PFL party as well the personal rivalry between ACM and Imbasashy also
Municipal revenues come from three main sources: (1) State government transfers amounting to 25% of the
revenues collected at the state level from the ICMS (consumption tax); (2) The ISS, which is a Municipal tax on
services; (3) The IPTU, which is a Municipal property tax. Because Bahia's population is poor, consumption is
limited, the revenues collected at the State level are limited, and transfers to the Municipal Government and
therefore ICMS contribution to Municipal revenues is low. Additionally, Salvador's economy is based on industry
rather than on services, and therefore revenues collected from the ISS are also limited. Finally, Salvador is a city
where property is mostly informal and revenues from the property tax are also very limited.
5 The revenues form transfers increased from R$196 in 1996 to R$276 in 1997.
52
played a major role in the exclusion of the Municipality from the Ribeira Azul project. 54
Partially because of the limited financial resources available to the Municipality, ACM and
Mayor Imbassahy (1996-2004) agreed that the Municipality would not be involved in the Ribeira
Azul project. Additionally, also as a consequence of the limited financial resources available at
the Municipal level, the PFL thought it more strategic from a political standpoint to spread the
available resources to construction or other type of projects in dissident areas of the city of
Salvador, so that the PFL could enhance its political influence over areas where there was not a
strong PFL contingency. However, despite this diversification strategy, in the summer of 2004,
during the political campaign for the Municipal elections in October 2004, CONDER's projects
have been extensively used by the PFL candidate Joio Enrique. For example, the face of Joio
Enrique appeared in advertisements of the Ribeira Azul project and other CONDER projects on
the local TV channels Bahia Globo and Itapoan.
Furthermore, the personal rivalry between ACM and Imbassahy, which developed over
time during Imbassahy's first and second mayoral administrations (1996-2004), is another
explanation for the exclusion of the Municipality. Mayor Imbassahy is a popular and respected
personality in Salvador. He succeeded at improving safety and renovating the historical center
of Salvador and the 'Dique do Tororo Park', thus making the city more attractive to tourists.
While ACM has a high opinion of Imbassahy, ACM feels threatened by Imbassay's leadership,
The relevance of party politics in influencing which agencies are involved in the Ribeira Azul project should not
be underestimated: political competition between two factions of the PFL, one led by ACM, controlling CONDER,
and the other one led by the current state governor Paolo Souto, controlling SEDUR, is currently causing major
conflicts between CONDER and SEDUR. CONDER and SEDUR are fighting for the control of financial resources
and responsibilities in the new $160 million government project Viver Melhor II. While CONDER is controlled by
ACM, SEDUR is controlled by Paulo Souto, and governed by Mousallen, Paolo Souto's right hand man. Such
competition between the two factions of the same party is motivated by the uncertainty over the name of the PFL
candidate who will run in the 2006 Governor elections. Since ACM's power and control over Bahia is decreasing
(Neto 2004), it is not obvious that PFL's new candidate for state governor will be chosen by ACM. Therefore, there
is tough political competition between the two factions.
5
autonomy, and success.
Since CONDER has been always controlled by ACM,
even after
other PFL governors succeeded him5 7 , he would not let the Municipality to participate to such an
important project. Summing up, even after the PFL gained control over the Municipality, the
Municipality surprisingly remained excluded from the Ribeira Azul project. Party politics, and
namely strategic decisions of the PFL party's leaders and personal rivalry within the PFL,
explain such exclusion.
3.5 Positive aspects associated with the exclusionof the Municipality
While the exclusion of the Municipality represents a major challenge for the
sustainability over time of the infrastructural improvements brought by the project, it has also
facilitated project implementation in three respects. First, as a result of the exclusion of the
Municipality, the project is insulated from potential intergovernmental political conflicts
between the State and Municipal government. This is especially important for projects - such as
the Ribeira Azul Projects - that cannot be completed within an election cycle. Municipal or state
elections can bring instability, if the Mayor and the governor are required to collaborate. In this
case, instead, political commitment at the state level is sufficient to continue construction. Note
that such insulation could be particularly important to complete the Ribeira Azul Project
especially after opposition candidates won the 2004 Municipal elections. This, however, does
not solve the long-run problem of maintaining the project. Second, it facilitated and accelerated
project implementation, in that responsibilities and tasks are not divided among different
agencies, and therefore the need for coordination is reduced. Additionally, such exclusion
" Interview with an important Bahian political scientist, who was part of the staff in the last ACM's administration
of the state government (1991-1995).
56 An example of ACM's control over CONDER lies in ACM's appointment of
CONDER's presidents.
57 Paulo Souto (PFL) between 1995 and 1999 and between 2002 and 2006 and Ceasar Borges (PFL) between 1999
and 2002,
maintained the focus remains only on the politically easiest task, which is construction, thus
allowing major infrastructure improvements to take place. Third, CONDER's exclusive
involvement in the project prevents it from blaming other implementing agencies for its
mistakes, or delays, thus creating incentives for CONDER to perform better. For example,
CONDER could not blame any other agencies for not having completed construction of a road in
Novos Alagados II.
3.6 Conclusion
This chapter described the relevance of politics - namely, the inter-governmental political
competition between the State and the Municipal government and party politics - in shaping
project design, and preventing the inclusion of the Municipality in the Ribeira Azul project.
Specifically, the chapter explained why the Municipality remained excluded at the beginning of
the Ribeira Azul project, and illustrates why, surprisingly, it remained excluded even once the
Municipality and the State government have been controlled by the same party. Finally, the
chapter presents some positive aspects that could derive from the exclusion of the Municipality
from project implementation. Namely these factors are: insulation from potential intergovernmental political conflicts between the state and the Municipal government, faster project
implementation because of the reduced need for inter-agency cooperation, and greater
commitment to the project because of the increased sense of ownership originated in the
exclusivity.
Chapter 4
Getting financial resources and becoming exemplary
4.1 Introduction
Previous chapters discussed important shortcomings of the Ribeira Azul project. This
chapter, however, explores that project's success, and the reasons why it has become well known
as an exemplary slum upgrading initiative. Section 4.2 discusses how the project succeeded in
securing financial resources from the Italian Government and the World Bank, which allowed
the project to be progressively scaled up to become a model for a state-wide upgrading initiative.
Section 4.3 explores the role marketing played in increasing the visibility of the project among
international donors and in promoting the project's reputation as a success. Section 4.4 describes
how the Ribeira Azul Project became exemplary. Section 4.5 concludes.
4.2 Securing financialresources
International donors and scholars often cite lack of financial resources as a reason for the
lack of sustainability among slum upgrading initiatives, many of which are delayed or benefit
only a limited number of slum dwellers. However, the Ribeira Azul program succeeded in
securing sufficient financial resources not only to complete the project, but to increase its scale to
the state-wide level.
The connections between the Director of the Italian Cooperation at the time, Vincenzo
Petrone, and AVSI's president Alberto Piatti, and the close relationship between the current
World Bank task manager, Ivo Imparato, and AVSI and CONDER, played a major role in
delivering financial resources to the Ribeira Azul program. Indeed, the close relationship
58
between Petrone and AVSI resulted in the quick approval of a $5 million grant from the Italian
government to finance the Technical and Social Assistance Project (PATS). PATS aims at
providing technical support for the implementation of the Ribeira Azul Program, assisting in the
scaling-up exercise in which the project moved from a small-scale initiative in Novos Alagados
to a larger program reaching across Ribeira Azul.59
Furthermore, Imparato's involvement in the Ribeira Azul project before he became task
manager of the World Bank during the preparatory phase of the PATS created a strong
connection between the World Bank and the institutions involved in Ribeira Azul. These
connections resulted in the World Bank's approval of a $96 million loan to scale up the Ribeira
Azul project to the state-wide level. The new government project, Viver Melhor II, will begin in
2006, and will involve upgrading projects in 16 Municipalities of the State of Bahia.
In order to understand how these personal connections were developed, and the
importance they played for delivering financial resources to the project, I will first analyze the
relationship between the Italian Ambassador Petrone, and AVSI's President Alberto Piatti. I then
discuss Imparato's career before he became task manager of the World Bank.
Inter-Italian Connections
The decision of the Italian government to provide the $5 million grant to the State
Government of Bahia is the result of a chain of events beginning in 1999, the main actors of
The project was presented for approval by AVSI in June, and was approved by the Italian government in July
2001.
59 The actions of the Technical and Social Support Project include: the coordination and assurance of
methodological consistency of the whole program; technical support for the elaboration of the physical and social
projects; support to Government actions in the area of housing; execution of strategic social work; systematization
and dissemination of the methodology to be employed; and part of the studies that will be necessary for the
preparation of Viver Melhor II.
58
which include: Vincenzo Petrone (Director of the Italian Development Cooperation at the time),
and AVSI's president Alberto Piatti. In 1999, when Petrone was still Director of the Italian
Development Cooperation, Piatti invited him to Salvador to visit the Ribeira Azul project.
Impressed by the physical improvements and the number of schools and training centers built in
Novos Alagados I Etapa,Petrone suggested that Piatti submit a proposal to the Urban Poverty
Reduction Fund of the Italian government60 in order to extend the project to reach the whole
Ribeira Azul area. AVSI subsequently submitted the Technical and Social Assistance Program
(PATS) proposal. As a result of Petrone's support, the Italian government approved the grant to
the state government of Bahia in July 2000, just one month after the request.
Petrone's active support to the project since its inception has two motivations. First, his
personal friendship with Alberto Piatti, which stimulated him to visit the project in 1999 and
follow the project through the approval process. According to AVSI's management and staff,
Petrone's support "was fundamental to make the social and technical assistant program possible"
and "without Petrone, this project would not have been in place or its approval would have taken
much longer." 61 Furthermore, it is important to note that, beside the being a friend, Piatti is also
an important political ally to Petrone in the Italian political arena; he is one of the most
influential members of "Comunione e Liberazione" (CL) 62 and "Compagnia delle Opere"
This fund was created by the Italian Government for financing Urban Poverty Reduction Initiatives, when the
Italian Government became a member of Cities Alliance.
61 Declaration of Alberto Piatti during a meeting in Salvador in
July 2004.
62 CL is an international ecclesiastic movement, born in Milan, and present in about seventy countries throughout the
world, with more than 100,000 members only in Italy. The purpose of this movement is "the education to Christian
62
maturity of its adherents and collaboration in the mission of the Church in all the spheres of contemporary life."
This includes the political sphere, an example being the governor of Lombardia. Additionally, CL is currently
supporting Berlusconi's party and his conservative coalition. The proximity between AVSI and CL is showed by
the fact that CL's members - who belong to the Fraternity of CL - the group of CL recognized by the Catholic
church are required to donate a monthly contribute to AVSI, the amount of which is voluntary.
60
(CDO)63, two organizations with major economic and political power in Italy. Members of CL
and CDO occupy important political positions in the Italian government and the private sector,
and wield considerable economic and political power in Italy. For example, the governor of
Lombardia, one of the richest and biggest regions in Italy is one of the leaders of CL. Second,
Petrone's active support to the PATS stems from his personal interest in promoting projects in
Brazil. In fact, when the negotiations for the approval of the Italian Grant started, Petrone was
aware that, starting from the year 2000, he was going to served as Italian Ambassador to Brazil.
In sum, the close relationship between AVSI and Petrone, and Petrone's interest in promoting
projects in Brazil, "delivered" the grant to the Ribeira Azul project.
Connections with the World Bank
The participation of the current World Bank task manager of the Ribeira Azul project,
Ivo Imparato, in the preparatory phase of the Social and Technical Assistance program (PATS),
created a strong connection between the World Bank and the Ribeira Azul project. This link
resulted in the World Bank's approval of a $96 million loan to the state government of Bahia
that, starting in 2006, will be used to scale up the operations of the Ribeira Azul project to reach
16 new Municipalities across the State of Bahia.
In order to understand how the direct link between the World Bank and the Ribeira Azul
program originated, it is necessary to look closer at Imparato's career before he became Task
Compagnia delle Opere (CDO), is an enterprise association with more than 30,000 members in Italy,63
and offers technical assistance on business development and marketing strategy, stimulating mutual collaboration
among its members (who are persons, as well as small, medium and large enterprises and non-profit organizations)
(CDO). 63 Thus, CDO is a network of firms and individuals, who have a catholic background, and a conspicuous
economic power. The extended network of relationships and resources offered by CDO aims - among other goals at developing constant relationships with Italian and International institutions, and at exerting political influence in
the Italian political arena (CDO). AVSI is particularly close to CDO, given that AVSI's current president Alberto
Piatti was in Board of Directors of CDO between 1997 and 2000, President of CDO in 1999, and vice-president of
CDO in 2003 and 2004.
63
Manager of the Ribeira Azul project in 2001. Imparato started working in Alagados for the first
time during the implementation of Novos Alagados I Etapa. At the time, as a staff member of
the Technical Central Unit at the Italian Minister for Foreign Affairs in Rome, he supervised the
pilot project in the small area of Nova Esperanza in Novos Alagados. Subsequently, Imparato
worked as a consultant for the Brazilian firm DiagonalUrbana,which the project subcontracted
to carry out socio-economic assessments in the communities of Alagados III and IV.
In 2000, during the preparatory phase of PATS, the Cities Alliance contacted Imparato,
and the Italian Government6 4 agreed to pay him to work in Salvador with AVSI and the Bahian
Government to define the institutional arrangement of the project and the social activities
included in the PATS. Imparato seemed the ideal person to carry out this work in Salvador in
direct contact with AVSI and the government of Bahia: he is Italo-Brazilian and fluent in Italian
and Portuguese, had previous experience working with the Italian Government, and knew the
people working on the project at the Italian Minister for Foreign Affairs. Additionally, because
of his experience supervising the project in Novos Alagados while working at the Italian
Minister for Foreign Affairs, and because of his consultancy for Diagonal Urbana, Imparato had
knowledge of and was committed to the project. During the preparatory phase of PATS,
Imparato interacted closely with AVSI's management and staff, preparing AVSI to comply with
65
the administrative and procurement requirements of the World Bank and Italian Government.
As a resulted, Imparato established a close personal relationship with AVSI's management.
When the preparatory phase ended, the World Bank hired Imparato as Task manager of the
Ribeira Azul project.
* Unfortunately, I was not able to understand who took the decision to employ him, whether it was the World Bank
or the Italian government, as I was given conflicting versions of the story.
65 Because the Italian Grant is channel via Cities Alliance, the administrative and procurement requirements are the
same as the World Bank. Therefore AVSI had to learn how to comply with those requirements.
Because of his close relationship with AVSI and - more limitedly - with CONDER,
Imparato was able as Task Manager to facilitate a close relationship beteeen the World Bank and
the various institutions involved in the Ribeira Azul project.. As early as 2002, only few months
after PATS began and Imparato became task manager, the World Bank - prompted by Imparato started planning Viver Melhor II which promises to be one of the most important state
66
government programs in the coming years. Also in 2002, the Italian government committed to
providing the government of Bahia with another grant, of 66 million.
Note that beside playing a fundamental role in delivering financial resources to the
project, the close connections between Imparato, AVSI and the Italian Government, as well as
the personal relationship between the Italian Ambassador to Brazil, Vincenzo Petrone, and
AVSI's president Piatti, created a strong alliance between AVSI, the World Bank, and the Italian
Government that had a major impact on the institutional arrangement of the project. In fact, as a
result of these connections, the Italian Government - supported by the World Bank and Cities
Alliance - decided to channel the Italian Grant through AVSI rather than CONDER, despite the
fact that the grant is intended to be for the State government of Bahia. 67 AVSI therefore occupies
a very powerful position in the project, which is reflected in the institutional arrangement of
PATS. Although AVSI only provides 3% of the resources ($2 million out of $60 million), or
12% if we consider the Italian Grant, AVSI co-directs the project.
4.3 Marketing
In addition to securing the financing and support of the World Bank and Italian
Government, the Ribeira Azul project has been very successful in marketing itself. Specifically,
In March 2005, the World Bank finally approved the disbursement of a $96 million loan to finance Viver Melhor
II, extending the Ribeira Azul project and methodology to another 16 Municipalities of the state of Bahia.
67 Interestingly, the public officials currently working in CONDER did not express any negative opinion about the
Italian government's decision to channel the grant through AVSI. However, as we have seen in chapter 2, the
decision of the Italian government faced the opposition of the Medici and his team.
66
AVSI's marketing strategy targeting international donors has been fundamentally important to
the project in two respects. First, it increased high level officials and workers' commitment to the
project. Second, it made international donors aware of the project, contributing to its fame as an
exemplary model of slum up-grading.
AVSI's marketing strategy
AVSI's marketing strategy consists of a number of initiatives in Italy and in the
international arena, aiming at disseminating the experience of the Ribeira Azul project, its
participatory approach, and its methodology. In Italy, AVSI has presented the project three
times at the Annual Meeting of the Catholic Movement "Community and Liberation" as an
example of catholic commitment to social development. This Meeting is an important social,
religious and political event that takes place annually in Rimini, Italy. Italian and international
politicians, economists, public and private managers and citizens' groups participate in the
meeting. The objective of this meeting is to generate discussion around different topics such as
ethics, international development, and political reforms. In all three occasions in which AVSI
presented the Ribeira Azul project, AVSI invited members of the government of Bahia to attend
the meeting. In 1997, AVSI invited Sonia Fontes, CONDER's president at the time, in 2000, the
newly elected CONDER's president Mario Gordilho, and in 2004 Mousallen, the director of
SEDUR. 68
AVSI's marketing strategy in the international arena includes the following components:
participation in international conferences, organizing seminars in Salvador and abroad, and
SEDUR is the Secretariat for Urban Development of the State of Bahia, and it is the agency from which
CONDER depends. The invitation of Mousallen to the Meeting, despite CONDER being in political competition
with SEDUR, is motivated by the fact that SEDUR will be the decision making institution, responsible for planning,
overall supervision and control of all Project activities and monitoring and evaluation in the next government Project
Viver Melhor II. Therefore, AVSI has a vested interest in collaborating with SEDUR.
68
activities to disseminate the methodology used by the program. First, AVSI attends international
conferences with the purpose of presenting the project and expanding networks with other
organizations, In 1996, AVSI presented the project to the Habitat Conference in Istambul.
Because it is currently running around 100 projects in 35 different countries and has good
reputation with international donors, AVSI is often asked to present its experience in
international conferences.
Second, AVSI organizes international seminars in Salvador and abroad. For example,
while I was in Salvador, AVSI organized a seminar on decentralization to which it invited the
president of AVSI, the president of SEDUR, CONDER's director of the Ribeira Azul project,
and the Italian Ambassador to Brazil. In 2000, AVSI organized an international seminar on
cooperation in the state of San Marino, and invited the current president of CONDER Gordilho.
Third, AVSI organizes activities to publicize the Ribeira Azul Project, and more
specifically AVSI's methodology. The methodology consists of an integrated, participatory and
cross-cutting approach (PATS Implementation Plan 200 1).69 These activities include
international seminars that specifically aim at presenting, discussing and disseminating the
Ribeira Azul methodology, developing and publicizing a bilingual manual (Portuguese/English)
which contains the systematization of the methodology, and producing a video and a CD-ROM
for its dissemination (PATS Implementation Plan 2001). Note that so-called "integrated"
projects are often characterized by the integration of physical improvements with social
activities. Thus, it is not particularly surprising that AVSI also used an integrated methodology,
and despite AVSI's officials focus on disseminating the characteristics, the innovative content of
AVSI define its methodology as (1) integratedbecause it seeks to include all aspects, physical as well as social,
that are part of the intervention; (2) participatorybecause it considers all stakeholders in the process as participants
in the planning and execution of the actions; (3) is cross-cutting because it attempts to reflect the importance of
elements such as gender and social capital, which are present in every aspect of the intervention (PATS 2001).
69
this methodology seems limited. However, due to astute marketing, AVSI's methodology is
considered by high level officials of AVSI and CONDER and of the Italian Government and
World Bank to be a fundamental and innovative contributing to the success of the Ribeira Azul
project. Indeed, donors and project stakeholders believe that infrastructure and housing
improvements could not have been implemented without AVSI's daily presence in the
community and the provision of training courses, environmental talks, and technical and
administrative support for community organizations.
Fourth, AVSI invites many Italian and international politicians, journalists and students
to visit the project every year. 70 These visits are important because the infrastructure and
housing improvement are quite impressive when viewed in person, especially when compared to
the conditions of those who have not benefited from the project yet. As shown in the pictures
contained in appendix I and III, there is a major difference between the environmental conditions
of the areas where the houses on stilts are still present, and the areas where houses were built.
Because many governments fear that NGOs will take all the credit for the project, they
often refuse to cooperate with NGOs. Interestingly, part of AVSI's marketing strategy is to
present the project as a "government project", thus granting ownership of the project to the
government. This approach is different than what I would have expected in that, while the
project is indeed a government project and AVSI contributes only with limited funding, AVSI
could, if it wanted, claim ownership of the project in the international arena because of its
In 2002, the project was visited by the following Italian top level politicians and Brazilian politicians: Baccini
(undersecretary of the Italian Minister for Foreign Affairs), Casini of the Italian Parliament, the president of AVSI
Piatti, a mission of Cities Alliance, a mission of the World Bank, the Governor of Bahia Paolo Souto (2002-2006),
the Major of Salvador Imbassay (1998-2004). Only while I was in Salvador from June to August 2004, the project
was visited by the following people: the Italian Ambassador in Brazil Petrone, two journalists of the Italian
cooperation writing a story about the project, 30 students from the most prestigious Italian University, AVSI's
president, the president of the Italian Catholic industrial association "Compagnia delle Opere," a World Bank
evaluator and the Task Manager for a week, 2 graduate students from the US doing field work on the project, a
journalist and a photographer working for the World Bank.
70
contacts. Additionally, CONDER has limited access international contacts, mostly, as we have
seen, because of the language barriers. I believe that presenting the Ribeiras Azul as a
government project is in AVSI's self interest in two respects. First, it prevents conflicts between
AVSI and the government of Bahia, since AVSI does not attempt to take all the credit for the
project. Second, the World Bank, Cities Alliance, and other donors express increasing interest in
promoting partnership between local government and NGOs, and therefore making it convenient
for AVSI to present itself as a partner of the government. In this way, AVSI, and ultimately
CONDER, have access to World Bank and Cities Alliance funding that are specifically set aside
to develop urban upgrading initiatives that foster collaboration between local governments and
civil society actors. Note that this does not mean that the World Bank, Cities Alliance, or the
Italian Government do not give AVSI credit for the project. For example, the World Bank task
manager of the Ribeira Azul Project, the task manager of the PRODUR project, and the person in
charge of the project in the Italian Minister for Foreign Affairs all underline how the success of
the project "would have not been possible without AVSI." 7 ' Before analyzing the effects of
AVSI's marketing strategy and its relevance for the success of the project, I explain how AVSI
created the international network that it used to publicize the project and the methodology.
Origins of AVSI's capacity to market the project
AVSI's capacity to market the project in the international arena is the result of many
years of experience working with international donors, in which AVSI developed contacts and
skills to interact with them, and built its reputation as an effective NGO. Additionally, because of
the many projects in different areas of the world and different sectors, AVSI is often invited to
international conferences to present its experience. As mentioned in chapter 1, AVSI has over
" Note that in session 3.4.3 we will present the different reasons that the different actors affirm to be the most
important contribution of AVSI to the project.
30 years of experience in designing and implementing projects financed by international donors.
In addition, AVSI is currently involved in more than 100 projects in 35 different countries,
financed by different donors -- including the World Bank, the Inter-American Development
Bank, the Italian Government, the European Union, US Agency for International Development
(USAID) and British Department For International Development (DFID). Moreover, AVSI
holds General Consultative status with the UN Economic and Social Council (ECOSOC) in New
York and the UN Industrial Development Organization (UNIDO) in Vienna,7 2 and is recognized
on the NGOs Special List of the International Labor Organization (ILO) in Geneva.
Furthermore, through the regular contact with international donors, AVSI built capacity
to interact with donors in both a literal and a "figurative" way. Such capacity grants AVSI the
possibility to play an intermediary role between CONDER and the international community.
According to CONDER's high level officials "AVSI is much more prepared than CONDER to
frame issues in the way donor agencies want, and knows what donors want to hear."73 My
interviews with AVSI's director in Bahia reinforce this evidence, in that he often referred to
concepts such as "bottom up approach to development," "community participation," and
"partnership" that are typically used by international donors. Furthermore, because of its
international contacts and language skills, AVSI is the institution which presents the project to
international visitors and politicians. Finally, note that AVSI's capability to interact with
international donors in a literal and figurative way also results in AVSI's workers taking over the
The consultative status allows NGOs to provide inputs to the ECOSOC into its deliberations at public meetings
and in its subsidiary bodies as well as in UN international conferences and their preparatory bodies. Each year the
approximately 2000 NGOs now holding consultative status receive the provisional agenda of ECOSOC. They have
certain privileges to place items on the agenda of ECOSOC and its subsidiary bodies; they may attend meetings,
where they may submit written statements and make oral presentations to governments
(http://www.ngocongo.org/ngopart/constat.htm).
73 Interviews with high level CONDER's officials who participated to the meetings of the World Bank, AVSI and
CONDER.
72
responsibility for writing project documents that should be written by CONDER's workers or by
workers of the two agencies together.7 4
Increasing high level officials and workers' commitment to the project
AVSI's marketing strategy had two important effects. First, it increased high level
officials' and workers' commitment to the project. Second, it provided the project with visibility
at the international level, giving it the reputation of an exemplary slum upgrading initiative. I
will now discuss the first one, whereas I will discuss the second one in the next session specifically focusing on how marketing contributed to make the project known as exemplary.
Governments often do not find upgrading attractive for political display, because the
upgraded areas are of too low a standard to be good "show pieces" (Otiso 2003). However,
AVSI's marketing strategy succeeded at increasing both high level officials' and workers'
commitment to the project. High level officials benefit personally and politically from the
greater international visibility of the project and connections with World Bank officials and
Italian politicians. CONDER's president Gordilho, as well as the director of the Ribeira Azul
project, have had the opportunity to travel to Washington, DC and Rome. Additionally, as
mentioned above, AVSI invited three Bahian government officials' to the Annual Meeting of the
CL in Italy. These invitations were important in increasing political commitment to the project
because they provided Bahian politicians with an understanding of the economical and political
power of AVSI's supporters in Italy. The most striking example of the relevance of the
participation in the CL Annual Meeting to increase the commitment of Bahian politicians to the
project is represented by Gordilho's visit in 2000. When he started his term in 2000, Gordilho
74 According to CONDER's staff, AVSI is also the "master of project document writing," precisely because of its
many years of experience in working with international donors.
was skeptical about a greater involvement of CONDER in the Ribeira Azul project because he
preferred CONDER to be involved in infrastructure projects in richer area of the city of
Salvador. Gordilho is an engineer, who used to work in private sector, first in the construction
industry and then in banking, and he never had experience in managing projects in poor
communities. Moreover, he is a top-down manager, unwilling to delegate responsibilities.
However, as a result of his participation to the Annual Meeting, as well as an international
meeting that AVSI organized in the State of San Marino, Gordilho acknowledged that
CONDER's participation in the Ribeira Azul project represented a "window that is a unique
possibility," and committed to work from then on with AVSI to solve the problem of Alagados. 7 5
Additionally, CONDER and other high officials of the state government who are not in direct
contact with international institutions such as the World Bank perceive CONDER's president
Gordilho and the Director of the Ribeira Azul Project Fonseca as having higher status because of
their international contacts. Another example of the effects of marketing on increasing high
level officials' commitment to the project is showed by the feelings of pride and honor expressed
by the director of the Ribeira Azul project Fonseca during international meetings where he
presented the project and during CONDER's internal seminars.
Social workers and engineers of both CONDER and AVSI also increased their
commitment to the project as a result of the marketing strategy. Many expressed their feeling of
pride and described their dedication to their job as a result of being the staff of a project that is
known not only in Bahia but also worldwide, is financed by different international donors, and
visited by many important personalities. Thus, workers became more dedicated to their job. For
example, even a worker who considers herself underpaid for the work she does, and is not
The international meeting in San Marino saw the participation of the high level officials of the Italian Minister for
Foreign Affairs, international donors, including the World Bank and Cities Alliance, Italian and International
academics.
75
satisfied with her working environment, affirmed that "I do not change job because this project is
famous, and surrounded by prestige. I want to be part of the success, that is why I work hard and
want to do the best I can to improve the living conditions of the inhabitants of the area."
Additionally, she added, "I will not have problems in finding a job when the project is over
because everybody knows it."
4.4 Becoming an "exemplary" project
Two factors contributed to make the Ribeira Azul Program an exemplary project: (1)
AVSI's marketing strategy and (2) the set of indicators used by international donors such the
World Bank and Cities Alliance to evaluate projects.
The role of AVSI's marketing strategy
While so far we have discussed how marketing contributed to the success of the project
we did not touch upon how it contributed to make the project exemplary. AVSI's marketing
strategy contributed to the portrayal of the project as an exemplary model among international
donors and the Italian government, by helping to promote the international reputation of the
Ribeira Azul project while hiding some of its shortcomings. 76 While other projects that had
equally important shortcomings were considered failures, AVSI's marketing succeeded at
becoming exemplary despite the fact that community participation in project design is limited,
and that the infrastructure built by the project is not maintained. Further, the project is unable to
recover costs from beneficiaries. The fact that World Bank officials still consider this project to
be exemplary, despite these failings, is striking, particularly given that cost recovery is a central
Note that this is not to say that the project was not successful. The project was indeed successful at implementing
components such as resettlement, and inter-agency cooperation (at least between CONDER and AVSI) that are
generally particularly challenging in slum upgrading projects.
76
criteria for World Bank assessments of project success. 77 The question remains as to whether
AVSI's reluctance to talk about the project's short comings (in particular the problems of
sustainability with regards to maintenance) and the continual efforts to present the project as a
success is preventing AVSI and CONDER from learning from mistakes and making adjustments
in the future. For example, while AVSI is focusing on marketing the project, it has not carried
out any evaluations of the project, including of the social activities to understand which ones
have been more or less successful. In addition, the wide recognition of the success of the project
has not stimulated innovation or modifications of the methodology, which instead has been
systematized to be used in the next government projects. Finally, while I would have expected
that marketing increased the cost of poor performance, thus stimulating institutional learning,
this does not seem to be the case of the Ribeira Azul project. In fact, as we have seen, marketing
seems instead to succeed at hiding project shortcomings rather than helping correcting them.
The role of the set of indicators of success used
The second reason why the Ribeira Azul program became well known as an exemplary slum
upgrading project is related to the set of indicators of success that are used by the World Bank
and Cities Alliance to evaluate projects. Cities Alliance has among its primary objectives
"improving the living conditions of the urban poor by developing city-wide and nation-wide
slum upgrading projects" and "scaling up successful initiatives" to expand the impact of
successful projects (Cities Alliance Charter). The World Bank, on the other hand, evaluates
outcome according to three factors: the relevance of the intervention's objectives in relation to
country needs and institutional priorities, efficiency (the extent to which the objectives have been
- or are expected to be - achieved without using more resources than necessary) and efficacy (the
7
Interview with Christine Kissides, World Bank Official.
extent to which the developmental objectives have been -or are expected to be - achieved) (OED
Evaluation Approach). If one evaluates the Ribeira Azul Program according to Cities Alliance's
objectives and the World Bank's outcome indicators, this project can be truly considered
exemplary. In fact, it started only targeting 15,000 people, it scaled up to target 150,000, and it
secured financial resources to be again scaled up to become a state-wide initiative in 2006.
Additionally, as we have seen, the project has built infrastructure and housing, thus meeting the
efficacy indicator.
On the other hand, the project is less successful if another set of indicators of success is
used. For example, if one assesses project success by looking at the involvement of its
beneficiaries in decision making and project design, or the likelihood of the project to be
maintained, the Ribeira Azul project cannot be considered as successful. This suggests that
when looking at exemplary projects, one needs to understand according to which set of criteria
they are considered successful. Furthermore, it suggests that powerful players such as the World
Bank and Cities Alliance, and their definition of project success play a major role in creating and
disseminating a model. Summing up, this project became well know as exemplary because the
World Bank and Cities Alliance use a specific set of indicators of success according to which
this project truly represents a success, and this set of indicators do not include "participation of
the community" and "maintenance of the infrastructure built."
4.5 Conclusion
This chapter presented two activities in which the project is successful: raising financial
resources and marketing. Specifically, it showed that because of a close personal connection
between the Italian Government and AVSI - through the Director of the Italian Development
Cooperation and AVSI's president, the project succeeded at receiving a $5 million grant of the
Italian Government. Additionally, the close involvement of the current World bank task manager
of the Ribeira Azul program in the preparatory phase of the Social and Technical Assistance
program financed by the Italian government (PATS) created a solid link between the World
Bank and CONDER and AVSI. As a result, the World Bank will provide a $96 million loan to
finance a $160 million slum upgrading initiative that will include 16 Municipalities of the State
of Bahia, using the same methodology of the Ribeira Azul project. Moreover, AVSI's marketing
strategy targeting Italian politicians and international donors increased high level officials and
workers' commitment to the project. Furthermore, AVSI's marketing explained how - despite its
shortcoming - the Ribeira Azul project has been recognized as an exemplary project and how - if
one assesses the Ribeira Azul project using the indicators used by the World Bank and Cities
Alliance - this project can be considered truly successful.
Conclusions
This paper examined the Ribeira Azul slum upgrading project in order to understand how
it worked and draw some lessons for designing projects. This project is worth a closer look
because the World Bank and Cities Alliance considered it exemplary, and, as such, a model for
other projects. However, when I started looking at the project, I realized that the Ribeira Azul
presents two important shortcomings. First, the project is not successful in ensuring community
participation in project design and decision making. Second, as in many other infrastructure
projects, in Ribeira Azul infrastructure is not maintained properly, which is an indicator of the
project's limited sustainability. On the other hand, despite these shortcomings, the project has
been successful in delivering infrastructure and housing, securing financial resources to scale up,
and marketing its methodology and accomplishments among international donors. Finally, the
massive marketing realized by the NGO AVSI and the particular set of indicators that
international donors typically use to evaluate project performance contribute to explain why this
project is worldwide known as an exemplary project. There are seven main lessons that emerged
from my analysis of the Ribeira Azul project and I will present each of them in turn.
1. Community participationin housing design and decision making is not a prerequisite
for success
Community participation in project design and decision making is not always necessary
to successfully complete construction of houses and infrastructure improvements. Participation at least in so far as community members' influence on project design and decision making - is
limited in the Ribeira Azul project. Community members complained about their lack of
involvement in the planning phase of the infrastructure and housing improvements, as well as
their limited influence over the design of the houses. Nevertheless, the Ribeira Azul project was
successful at providing houses, building infrastructure such as drainage systems and roads, and
improving access to water, sanitation, and electricity. There are three five main reasons why
CONDER and AVSI succeeded at building housing and infrastructure. First, the poor living
conditions of the inhabitants of Alagados and Novos Alagados, and the high costs of
maintenance of the houses on stilts, motivated project beneficiaries to support the project.
Second, CONDER's decision to limit resettlement by land-filling the tidal area where stilt houses
were present also encouraged people living in those areas to allow their house to be replaced
because of the improvement in living conditions of on ground housing. Third, CONDER's
decision of stopping paying to maintain the walkway made of planks that connect houses on
stilts indirectly pressured the inhabitants to participate in the project. Fourth, CONDER's
decision not to collect payments for the houses also facilitated the beneficiaries' acceptance to
resettle or to receive a house they were not happy with. Finally, small changes to project design
that were introduced to accommodate community members' needs were also critical to speed up
construction.
2. NGOs are not necessarily representativeof the community
In Ribeira Azul project the presence of an NGO, AVSI, in the implementation of
infrastructure projects did not ensure that community members' needs were taken into account
-
at least for what regards housing design and decision making. This suggests that, at least in this
case, there is a myth about NGOs being close to the community and able to respond to
community members' concern that needs to be debunked. AVSI is not perceived by project
beneficiaries as an independent from government because of AVSI's close collaboration with
CONDER. Additionally, because AVSI did not want to jeopardize this close collaboration with
CONDER, it did not play the intermediary role between the community and CONDER that we
would expect from an NGO. Finally, my interviews suggest that there is a trade-off between
scaling up and ensuring community participation. In fact, when the project was smaller and
fewer donors were involved, project staff had more time to interact and a closer relationship with
community members.
3. Government can also have the capacity to customize solutions to communitv's needs
In contrast to the widespread belief that centralized infrastructure agency do not have the
experience or human resources to directly interact with community members, in this project the
state government agency CONDER was successful in customizing solutions to community's
needs. This was possible thanks to a group of socially-minded committed technicians working
for CONDER who developed a close relationship with the community.
4. "Engineersin need" work with social intermediationspecialists
While technical staff often refuses to cooperate with social intermediation professionals
or to take into account their technical inputs, this was not the case in the Ribeira Azul project. In
this project, engineers successfully cooperated with social intermediation specialists and took
into account their technical suggestions when social intermediation specialists made their work
easier. This is because these engineers had limited - if any - experience in working in poor
communities and therefore found it helpful to have social intermediation specialists dealing with
community members' complaints and concerns. Furthermore, informal interactions over lunch
breaks also created opportunities for exchange between engineers and social intermediation
specialists and increased engineers' trust and willingness to listen to social intermediation
specialists' suggestions.
5. Lack of coordinationwith or without politicalparty alignment
Inter-governmental political competition between different level of governments and
party politics were very powerful forces in shaping project design and implementation. More
specifically, inter-governmental competition between the state of Bahia and the Municipality of
Salvador caused the exclusion of the Municipality - which is the institution responsible for
maintenance - from the initial phase of design and implementation of the project. Even when the
Municipality and the state government came to be controlled by the same party - i.e. the PFL the Municipality was still excluded from playing an active role in the Ribeira Azul project. This
is striking because one would expect greater cooperation when two agencies are controlled by
the same political party. Such exclusions has two explanations related to party politics: (1) the
PFL party decided to strategically spread the limited resources available to the Municipality in
dissident areas of the city of Salvador rather than in the area of the Ribeira Azul project where
there already was a strong PFL contingency; (2) the leader of the PFL party, controlling
CONDER, and the mayor of Salvador, also from the PFL party, had a personal rivalry that
resulted in competition between the two organizations.
6. The important role of marketing
Marketing can play a very important role in increasing the commitment of high level
officials and workers to the project, and, ultimately, contributing to the success of the project.
Indeed, the NGO AVSI intensively promoted the project among international donors, a situation
that created the opportunity for project officials and politicians to establish direct contacts with
international donors. As a result of these contacts, high level officials benefited personally and
politically from a greater involvement in the project, which ultimately increased their
commitment to and interest in the project. Furthermore, marketing also increased workers'
commitment. In fact, project staff felt proud of being part of such a well known project and this
feeling resulted in greater dedication to their job. AVSI's marketing strategy included the
presentation of the project in international conferences, the systematization and dissemination of
the methodology through the publication of manuals, and the organization of seminars. In
addition, AVSI repetitively invited CONDER and state government high level officials to
international meetings in Italy. Similarly, many international politicians and authorities visit the
project in Brazil, providing more opportunity for higher level officials to interact with
international officials. However, it is important to note that marketing can have some drawbacks.
Specifically, marketing may have prevented AVSI from innovating and learning from past
mistakes. For example, while focusing on marketing, AVSI did not carry out any evaluation of
the project.
7. Success depends on the set of indicatorsused
This project is truly successful according to the set of indicators of success used by the
World Bank and Cities Alliance and namely "ability to scale up" and "ability to build
infrastructure." Instead, the project is much less successful if other indicators of success such as
"community participation in project design" or "sustainability of infrastructure improvements"
are used. Indeed, if one assesses the Ribeira Azul project using these last indicators, one will
find that the project is not particularly successful because, as I illustrated in the paper, the project
did not elicit sufficient community participation and the lack of involvement of the municipality
(the main actor in charge of maintenance) may jeopardize project sustainability. However,
because the World Bank and Cities Alliance use the first set of indicators to evaluate projects,
this project is portrayed as exemplary. This lesson is an indication of the important role that
powerful players in the development field, such as the World Bank, Cities Alliance, and their
definition of project success, play in creating and disseminating models for slum upgrading.
APPENDIX I: Neighborhoods of Alagados and Novos Alagados
Picture: Texeira (2002)
APPENDIX II: Comparison between Novos Alagados before
I
and after the intervention
Novos Alagados I and II Etapa:
1984 and 2003
Source: AVSI
On the left hand side, the tidal zone is occupied by the houses on stilts. On the right hand side,
the houses on stilts have been removed and substituted with houses. Additionally, a costal road
has been built.
APPENDIX III: Houses on stilts
A) Pier
b) View of the houses on stilts from the bay. As you can see, the water is full of trash.
c) House on stilts
APPENDIX IV: Houses before and after the intervention
On the left hand side, pictures of the houses on stilts after the land-fill has been done, but before
they have been replaced by new houses. On the right hand side, the costal road, and some
examples of the houses provided by the project.
Bibliography
Bamberger, Micheal, Sanyal, Bish, and Nelson Valverde (1982). Evaluation of sites and
Services Projects - The Experience from Lusaka, Zambia. World Bank Staff Working Papers
Number 548
Blackburn, James and Jeremy Holland (1998)._Who Changes - InstitutionalizingParticipationin
Development. London UK: Intermediate Technology Publications Ldt.
Boris, Elizabeth, and Eugene Steurerle (eds.) (1999). Non Profits and Government Collaborationand Conflict. Washington DC: The Urban Institute Press.
Brinkerhoff, Jennifer (2003). Donor Funded Government-NGO partnership for public service
improvement: cases from India and Pakistan. Voluntas 14 (1).
Cernea, Michael (1988). Nongovernmental Organizations and Local Development." Washington
DC: The World Bank. World Bank DiscussionPapersno. 40.
Chisholm, Donald (1989). Coordinationwithout Hierarchy- Informal Structures in
MultiorganizationalSystems. University of California Press, Berkley.
Cities Alliance (2004). Annual Report. Washington DC: Cities Alliance.
Clark, John (1995). The State, Popular Participation and the Voluntary Sector. World
Development 23 (4): 593-601.
Davis, Jennifer (2001). Seeking Good Governance: Experiments from South Asia's Water &
Sanitation Sector. Unpublished
Davis, Jennifer (2002). Taking Urban Upgrading to scale: where are the bottlenecks? Retrieved
on may 13, 2004 from
http://web.mit.edu/urbanupgrading/waterandsanitation/action/Davisupgrading.pdf
DiMaggio, Paul and Walter Powell (1982). The Iron Cage Revisited: Conformity and Diversity
in OrganizationalFields. PONPO Working paper 52. Retrieved February 2, 2005, from
http://ponpo.som.yale.edu/work.php
De Carvalho, Jose' Murilo (2001). Cidadaniano Brasil. Rio de Janeiro: Civilizagao Brasileira.
Ebrahim, Alnoor (2003). Accountability in practice: Mechanisms for NGOs. World
Development 31 (5), 813-829.
Evans, Peter (1996). Government Action, Social Capital, and Development: Reviewing the
Evidence of Synergy. World Development, 24 (6), 1119-1132.
Fernandes Arau'jo Antonio Sergio (2004). Gestao Municipal e Participacaosocial no Brasil: a
trajetoriade Recife e Salvador (1986-2000). Annablume; Fapesp, Sao Paolo.
Fiszbein, Ariel and Pamela Lowden (1999). Working Toghether for a Change - Government,
Civic and Business Partnerships for Poverty reduction in Latin America and the Caribbean. EDI
LearningResources Series. Economic Development Institute. Washington DC: World Bank
Fowler, Alan (2000). Ngo futures: beyond aid: NGDO values and the fourth position. Third
World Quarterly,21 (4), 589-603.
Fyvie, Claire and Ager, Alastair (1999). NGOs and Innovation: Organizational Characteristics
and Constraints in Development Assistance Work in the Gambia. World Development, Vol. 27,
No. 8, pp.1383-1395.
Gibbs, Christopher; Fumo, Claudia, and Thomas Kuby (1999). Non governmental organizations
in World Bank - supportedprojects: a review. World Bank Operations and Evaluations
Department. Washington DC: The World Bank.
Gordilho Souza, Angela (2000). Limites do Habitar- Segregacao e exclusao na configuracao
urbana contemporaneade Salvador e perspectivas no final do seculo XX. Salvador da Bahia,
Brazil: Editora da Universidade Federal da Bahia (EDUFBA)
Governo do Estado da Bahia- SEPLANTEC/CONDER, Associacao Voluntarios para o Servico
Internacional, Fundacao Don Avelar Brandao Vilela (1993). Novos Alagados, Recuperacao
Ambiental e Promocao Social. Projeto de Cooperacao Tecnica Internacional Brasil-Italia, Plano
Operativo, Salvador - Bahia.
Hirschman, Albert (1967). Development Projects Observed. The Brookings Institution.
Washington DC.
Huxham, Chris, and Siv Vangen (2000). Ambiguity, complexity and dynamics in the
membership of collaboration. Human Relations, Vol. 53 (6), pp. 771-806
Huxham, Chris, and Vangen Siv (1996). Working together: Key themes in the management of
relationships between public and non-profit organizations. InternationalJournalofpublic
Sector Management, Vol.9 (7).
Imparato, Ivo and Jeff, Ruster (2003). Slum Upgradingand Participation.Lessons from Latin
America. The World Bank.
Kalegaonkar, Archana, and David Brown (2000). IntersectoralCooperation:lessonsfor
practice,IDR reports, vol. 16 (2), Boston MA.
Kooiman, Jan (1993). Modern Governance. SAGE.
Lipsky, Micheal (1980). Street-Level Bureaucracy- Dilemmas of the Individual in Public
Services. Russell Sage Foundation, New York
McCarney, Patricia (1987). The Life of an Idea: the Rise and Fall of Sites and Services at the
World Bank. Master Thesis, Department of Urban Studies and Planning, MIT.
Medcof, John (1997). Why too many alliances end in Divorce. Long Range Planning,Vol.30
(5), pp. 718-732.
Melo, Marcus Andre (1999). Reforma do Estado e MudancaInstitucionalno Brasil. Fundacao
Joaquim Nabuco, Editora Massangana, Recife.
Meyer, Carrie (1995). Opportunism and NGOs: Entrepreneurship and Green North-South
Transfers. World Development, Vol. 23 (8), pp. 1277-1289.
Moore, Mick (2005). The Politics and Bureaucratics of Rural Public Works: Maharashtra's
Employment Guaranteed Scheme. Draft.
Nance, Earthea (2004). Putting Participation in Context: an Evaluation of Urban Sanitation in
Brazil. PhD Dissertation, Civil and Environment Engineering of Stanford University.
Ndegwa, Stephen (1996). The two faces of Civil Society: Ngos andPolitics in Africa. Kumarian
Press.
Neto, Paulo Fabio Dantas (2003). Surf nas Ondas Do Tempo: do Carlismo Hist6rico ao Carlismo
P6s-Carlista. Caderno CRH, no. 39. Centro de Recursos Humanos, UFBA 2003.
Operations Evaluation Department (1999). Nongovernmental Organizationsin World Bank-
supported Projects. Washington DC: The World Bank.
Operations Evaluations Department (2003). Improving the Lives of the Poor through Investment
in Cities: An update on the Performanceof the World Bank's Urban Portfolio. Washington DC:
The World Bank
Operations Evaluations Department (1997). Brazil-LearningfromBest Practicein Five urban
Projects:Medium -Sized cities, the Recife MetropoliticanRegion, Preparationsof Metropolitan
Development Programsfor Fortalezaand Salvador, the ParanaMarket Towns Improvement and
northeast Urban FloodReconstructionProject. Washington DC: The World Bank
Operations Evaluations Department (1996). Enhancing the Quality ofLife in Urban Indonesia.
Washington DC: The World Bank
Operations Evaluations Department (1992). IntegratedUrban Projects: Experience in Ecuador,
India, andBrazil. Washington DC: The World Bank
Operations Evaluations Department (1994). Two Decades ofLendingfor UrbanDevelopment.
Washington DC: The World Bank
Osborne, Stephen (Ed.) (2000). Public-Privatepartnerships- Theory andpractice in
InternationalPerspective.Routledge Advances in Management and Business Studies.
Osborne, Stephen (1997). Managing the Coordination of Social Services in the Mixed Economy
of Welfare: Competition, Cooperation or Common Causes? British Journalof Management, 8,
pp. 317-328.
Otiso, Kefa (2003). State, voluntary and private sector partnerships for slum upgrading and
basic service delivery in Nairobi City, Kenya. Cities, 20 (4), pp. 221-229.
Salomon, Lester (1995). Partnersin Public Service: Government-Nonprofit relations in the
modern Welfare state. John Hopkins University Press
Sanyal, Bish (1994). CooperativeAutonomy: the Dialectic of State-Ngos Relationship in
Developing Countries. Research Studies 100, International Institute for Labour Studies, Geneva.
Sanyal, Bish and Meenu Tewari (1991). Politics and Institutions in Urban Development - the
Story of the Calcutta Metropolitan Development Authority. In Asian Development Bank, The
Urban Poor andBasic InfrastructureServices in Asia and the Pacific.Vol III. Asian
Development Bank and Economic Development Institute.
Smith, Steven Rathgeb and Micheal Lipsky (1993). Nonprofitsfor Hire - the Welfare state in
the Age of Contracting.Harvard University Press.
Souza, Celina (1997). ConstitutionalEngineeringin Brazil - The Politics of Federalismand
Decentralization. London: Macmillan Press LTD and New York: St. Martin's Press INC.
Souza, Celina (1999). Decentralization, Local Autonomy and Metropolitan Influence: the case of
Salvador, Brazil. In Adrian Atkinsohn, Julio Davial, Edesio Feranndes, Michael Mattingly (eds.),
The Challenge ofEnvironmentalManagement in UrbanAreas. Ashgate. Publishing Ldt.
Souza, Celina (2004). Regioes Metropolitanas: Trajetoria e Influencia das Escholhas
Institucionais. In Queiroz Ribeiro Luis Cesar (eds.), Metropoles: entre a coesdo e a
fragmentagdo, a cooperagdo e o conflito. Editoria Fundagio Perseu Abramo. Sio Paolo.
Superintendencia de Estudos Economicos e Socias da Bahia (SEI) (2002). Estadisticados
Municipios Baianos. Salvador da Bahia, Brazil: SEI.
http://www.sei.ba.gov.br/publicacoes/bahia dados/est municipios/xls/2002/3.1.xls
Superintendencia de Estudos Economicos e Socias da Bahia (SEI) (2003). Bahia en Sintese.
Salvador da Bahia, Brazil: SEI. http://www.sei.ba.gov.br/bahia sintese/index.htm
Teixeira de Carvalho, Eduardo (2002). Alagados da Bahia - Intervencoes Publicase
ApropriacaoInformal do Espaco Urbano. Universidade Federal da Bahia, Faculdade de
Arquitetura, Programa de Pos-Graduacao em Arquitetura e Urbanismo, Mestrado em Arquitetura
e Urbanismo.
Tendler, Judith (1982). Turning Private Voluntary Organizationsinto Development Agencies:
QuestionsforEvaluation. AID Program Evaluation Discussion Paper No. 12
Tendler, Judith (1993). "Reinventing the Projects." Executive summary and chapter 2 in: New
Lessons From Old Projects: The Workings of Rural Development in Northeast Brazil,A World
Bank Operations Evaluation Study. Pp. xix-xxvi, 14-33.
Tendler, Judith (1994). The Blurredboundaries between public andprivate sector. Conference
Session on Social Capital and Economic Development.
Tendler, Judith (1997). Good Government in the Tropics. The Johns Hopkins University Press,
Baltimore and London
Tendler, Judith (1999). Thoughts for a research Agenda for the meeting on public-private
partnerships, University of Michigan, 25 June 1999. Unpublished
The World Bank (1992). Project Performance Audit Report, Brazil, Report no. 10832.
The World Bank (2002). "Non-governmental organizations and civil society engagement in
world bank supported projects: lessons from OED evaluations". OED Lessons and practices no.
18. The World Bank, Washington DC.
The World Bank (2005). Implementation Completion report on a Loan in the amount of $100
Million to the State of Bahiaforthe Bahia Municipal InfrastructureDevelopment and
ManagementProject. Retrieved on March, 2 4 th 2005 from http://www-
wds.worldbank.org/servlet/WDSContentServer/WDSP/IB/2005/04/01/000090341_20050401084
223/Rendered/PDF/31877.pdf
UNHABITAT (2003). The Challenge of Slums - GlobalReport on Human Settlements 2003.
UNHABITAT.
Van de Ven, Andrew and Diane Ferry (1980). Measuringand Accessing Organizations.New
York: Wiley.
Verma, Gita Dewan (2000). Indore's slum project: a worm's eye view from the ground. Habitat
International24, pp. 91-117
Vittadini, Giorgio (1997). Il Non-ProfitDimezzato. RCS libri & Grandi Opere S.p.A.
Weiss, Janet (1981). Substance vs. Symbol in Administrative Reform: the case of Human
service coordination, Policy Analysis, Vol 7.
100
Weiss, Janet (1987). Pathways to Cooperation among Public Agencies. JournalofPolicy
Analysis and Management, 7 (1), pp. 94-117.
Werlin, Herbert (1999). The Slum Upgrading Myth. Urban Studies, 36 (9), 1523-1534.
Wilson, James (1989). Bureaucracy:what government agencies do and why they do it. Basic
Books Inc.
101